@blamejs/exceptd-skills 0.19.15 → 0.19.16

package/data/zeroday-lessons.json

@@ -16039,7 +16039,29 @@
     },
     "ai_discovered_zeroday": false,
     "ai_discovery_source": "vendor_research",
-    "ai_assist_factor": "none"
+    "ai_assist_factor": "none",
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-001",
+        "name": "CISA-KEV-RESPONSE-SLA",
+        "description": "The asset on the clock for this CVE is not a server in the production estate: the packet puts the command execution on the developer host, reached through the React Native Community CLI's Metro development server, so the change-managed pipeline that patches production never sees the vulnerable component and the KEV clock has to be pointed at developer workstations and CI/build images explicitly. Requirement: from the 2026-02-05 KEV listing, every host and every build image that can start Metro reaches the fixed CLI release, with completion measured per host and per image rather than by a production deployment count. The packet registers no live-patch path and states the vendor patch typically requires a service restart or system reboot, so a Metro process still running from a pre-fix load remains exposed until it is stopped and restarted on the fixed release — a host where the package was updated while a dev server kept running is not remediated, and this is the specific way the completion count goes wrong on a machine a developer leaves running for days. Compensating control for the window before that completes, with its precondition stated: the packet's stated access requirement is a network attacker reaching the exposed dev server, so binding Metro to loopback and denying inbound connections to its port at the host firewall bounds who can send the POST. That lever exists only where the workflow does not need off-host reach — a physical test device, an emulator or simulator on another machine, or a teammate or CI runner loading the bundle all require the server to answer from the network, and on those hosts the restriction is simply unavailable. It also does not repair the endpoint: anything still permitted to reach the dev server reaches the command-injection path unchanged, and the packet records that on Windows the attacker additionally controls shell arguments in full. Reachability restriction is a holding measure for the window before the fixed release and its restart land, never the recorded remediation.",
+        "evidence": "Packet: CWE-78 OS command injection in the React Native Community CLI's Metro development server; unauthenticated network attackers send POST requests to the Metro Development Server and run arbitrary executables via a vulnerable endpoint exposed by the server, and on Windows can execute arbitrary shell commands with fully controlled arguments; the attack_vector places the execution on the developer host. cisa_kev true, kev_date 2026-02-05, active_exploitation confirmed, poc_available true, CVSS 9.8, RWEP 77. patch_available true; live_patch_available false with live_patch_notes recording no live-patch tool registered for this entry and a vendor patch that typically requires service restart or system reboot per the KEV requiredAction.",
+        "gap_closes": [
+          "NIST-800-53-SI-2",
+          "ISO-27001-2022-A.8.8"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-018",
+        "name": "SCANNER-PAPER-COMPLIANCE-TEST",
+        "description": "The verdict that goes wrong on this CVE is a developer-fleet scan reporting no vulnerable software present. What is vulnerable is a command-line development tool a developer runs inside a project, and the exposure exists only while that tool's Metro dev server is live and answering from the network — neither fact is visible to a scanner that enumerates installed products and OS patch level, so an estate can hold a clean report while a fully exploitable dev server is listening on the same LAN as the attacker. The operational test has two halves and both are required. First, enumerate every developer host and every CI/build image capable of starting Metro and show the React Native Community CLI each one resolves is at or above the fixed release, rather than showing that the workstation's operating system and registered applications are current. Second, with a project running on a representative developer host, send a POST to that host's Metro dev server port from a different host on the same network segment and confirm it is refused before any executable is invoked. The second half is the one the patch dashboard structurally cannot answer, and it is the half that corresponds to the packet's actual access requirement. Scope discipline matters here: the packet ties this CWE-78 sink to the React Native Community CLI's Metro development server and gives no basis for treating other frameworks' development servers as instances of this CVE, so the test population is hosts that run this CLI — widen it only where a verified source identifies another product carrying the same endpoint, because instructing operators to hunt every development server in the estate manufactures findings against tooling no evidence implicates. Limits: this control produces a truthful verdict, it does not remediate. The packet records a vendor patch with no live-patch path and a restart requirement, so hosts the test finds exposed still have to reach the fixed release and restart the dev server.",
+        "evidence": "Packet: the exploit requires unauthenticated network attackers to send POST requests to the Metro Development Server, an endpoint exposed by the React Native Community CLI on the developer host (CWE-78); on Windows the attacker also executes arbitrary shell commands with fully controlled arguments. cisa_kev true (2026-02-05), active_exploitation confirmed, poc_available true, CVSS 9.8, RWEP 77, patch_available true, live_patch_available false with the vendor patch typically requiring service restart or system reboot.",
+        "gap_closes": [
+          "UK-CAF-B4",
+          "NIS2-Art21-network-security"
+        ]
+      }
+    ]
   },
   "CVE-2026-24423": {
     "name": "SmarterTools SmarterMail Missing Authentication for Critical Function Vulnerability",
@@ -18957,7 +18979,31 @@
     },
     "ai_discovered_zeroday": false,
     "ai_discovery_source": "vendor_research",
-    "ai_assist_factor": "none"
+    "ai_assist_factor": "none",
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-030",
+        "name": "PERIMETER-DEVICE-ZERODAY-SLA-TIER",
+        "description": "ArrayOS AG is the operating system of a secure-access gateway, so the appliance the packet describes is not a server behind the perimeter — it is the perimeter, and the packet places an unauthenticated OS command-injection path on it with confirmed in-the-wild exploitation and a public exploit. That is the combination this tier exists for: a routine 14- or 30-day application patch window leaves the device that terminates untrusted remote access executing attacker-supplied commands for weeks, and no severity-band justification changes that, because the exposure is where the flaw sits rather than how it scores. For this CVE the clock runs from the 2025-12-08 KEV listing to the point where each Array Networks appliance running ArrayOS AG is on the vendor's fixed build and has taken the restart the fix requires; the packet registers no live-patch path and notes the vendor patch typically requires a service restart or system reboot per the KEV requiredAction, so an appliance that has staged the fixed image but has not restarted still runs the vulnerable code and must be counted as exposed rather than as patched-pending-window. Scope this to the ArrayOS AG appliances the packet names; it gives no basis for treating other gateway products in the estate as instances of this flaw. Precondition on the tier's alternative, which is where this control is usually over-claimed: restricting which source networks can reach the appliance's exposed interface bounds who can send the injection, it does not repair it, and a secure-access gateway exists precisely to answer clients on untrusted networks — for an appliance publishing remote access to the internet there may be no source restriction that removes the path at all, only a reduction to the population that must be able to reach it, every member of which still satisfies the packet's stated access requirement of unauthenticated network reach.",
+        "evidence": "Packet fields for CVE-2025-66644 (Array Networks ArrayOS AG OS Command Injection Vulnerability, CWE-78): CISA KEV listed 2025-12-08, active_exploitation confirmed, poc_available true, CVSS 9.8, rwep_score 77. The attack_vector describes an OS command-injection flaw enabling unauthenticated remote command execution on the secure-access gateway appliance; the vector states ArrayOS AG contains an OS command injection vulnerability that could allow an attacker to execute arbitrary commands. patch_available true, live_patch_available false, with live_patch_notes stating no live-patch tool is registered for this entry and that the vendor patch typically requires service restart or system reboot per the KEV requiredAction.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "NIS2-Art21-patch-management",
+          "NIST-800-53-SI-2",
+          "UK-CAF-B4"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-032",
+        "name": "PERIMETER-COMPROMISE-REBUILD-NOT-PATCH",
+        "description": "The packet states this control's trigger condition outright: unauthenticated remote command execution on a perimeter secure-access gateway, KEV-listed 2025-12-08, active exploitation confirmed and a public exploit available. For any Array Networks appliance running ArrayOS AG that was reachable during the exposure window, the default response is therefore not upgrade-in-place. Capture the configuration and whatever logs the box still holds for forensics, rebuild the appliance from vendor media onto the fixed build, and rotate everything it held or brokered — local administrative accounts, the directory or RADIUS bind credentials it authenticated remote users against, its TLS server keys, and any session or token material issued to users through it — because arbitrary command execution as the appliance gave the attacker the same read of that material that the appliance itself had. The vendor fix closes the injection path; it removes nothing already written through it, so an added administrative account, a modified startup script or an implant on the filesystem survives the upgrade and is inherited by the 'remediated' device. This is also the honest answer to the least-privilege gap recorded against this entry: the attacker never authenticates as any gateway operator, so per-account privilege scoping is never consulted and a least-privilege attestation passes cleanly while the path stays fully open — the only privilege containment available after the fact is the rotation, because the blast radius is defined by what the appliance's own identity could reach, not by any user account. The distinguishing test is a records test rather than a scan: for every gateway remediated since the KEV listing, ask for the rebuild evidence and the credential-rotation record alongside the build number. An estate that can produce only 'upgraded to the fixed build' has a patched appliance that may still be attacker-held. Precondition: the scope of this response is set by an exposure-window determination, and the packet supplies no compromise indicator to clear an individual appliance against — so where reachability during the window cannot be excluded, the appliance is treated as compromised rather than cleared on the absence of an alert, and the on-box logs used to make that call are themselves writable by an attacker who held command execution.",
+        "evidence": "Packet fields for CVE-2025-66644: CWE-78, cisa_kev true with kev_date 2025-12-08, active_exploitation confirmed, poc_available true, CVSS 9.8, rwep_score 77. The attack_vector records unauthenticated remote command execution on the secure-access gateway appliance. patch_available true and live_patch_available false, with live_patch_notes stating the vendor patch typically requires service restart or system reboot per the KEV requiredAction. NIST SP 800-53 AC-6 (Least Privilege) and ISO/IEC 27001:2022 A.8.8 (Management of technical vulnerabilities) are both recorded as citing framework gaps on this entry — the first is never consulted on an unauthenticated path, and the second closes at 'patch applied'.",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.8",
+          "NIST-800-53-AC-6"
+        ]
+      }
+    ]
   },
   "CVE-2025-55182": {
     "name": "Meta React Server Components Remote Code Execution Vulnerability",
@@ -19017,242 +19063,273 @@
     },
     "ai_discovered_zeroday": false,
     "ai_discovery_source": "vendor_research",
-    "ai_assist_factor": "none"
-  },
-  "CVE-2021-26828": {
-    "name": "OpenPLC ScadaBR Unrestricted Upload of File with Dangerous Type Vulnerability",
-    "lesson_date": "2026-05-29",
-    "attack_vector": {
-      "description": "an unrestricted file-upload flaw (CWE-434) on the ScadaBR SCADA/HMI web interface, letting an unauthenticated attacker upload a file (e.g. a web shell) for code execution on the HMI server. CISA KEV-listed 2025-12-03 with confirmed in-the-wild exploitation.",
-      "privileges_required": "none (the device is reachable by an unauthenticated attacker; exposure is amplified when the OT zone is not segmented)",
-      "complexity": "low — KEV-listed, actively exploited; treat as weaponized",
-      "ai_factor": "No AI involvement documented in discovery or weaponization."
-    },
-    "defense_chain": {
-      "prevention": {
-        "what_would_have_worked": "Apply the vendor firmware/update where one exists; where the device cannot be patched, isolate it in a segmented OT zone (zones-and-conduits / Purdue model), block all IT and internet reachability, and restrict access to authenticated engineering workstations.",
-        "was_this_required": true,
-        "framework_requiring_it": "CISA BOD 22-01 (KEV remediation) + IEC 62443-3-3",
-        "adequacy": "Patching is often impossible on OT; segmentation and access restriction are the real controls, and a flat or internet-exposed OT network defeats them."
-      },
-      "detection": {
-        "what_would_have_worked": "OT-network monitoring for unauthorized connections to the ScadaBR HMI, unexpected configuration/logic changes, and access from outside the device's intended zone.",
-        "was_this_required": false,
-        "framework_requiring_it": null,
-        "adequacy": "Necessary because unpatched OT devices may stay exploitable indefinitely; behavioral detection is the backstop."
-      },
-      "response": {
-        "what_would_have_worked": "Isolate the device, validate and restore its configuration/logic from a known-good baseline, rotate any credentials/keys it held, and investigate for safety-impacting manipulation; engage OT/safety engineering before any change.",
-        "was_this_required": true,
-        "framework_requiring_it": "NIST 800-82r3 + NIST 800-53 IR-4",
-        "adequacy": "Mandatory; OT compromise can have physical/safety consequences, so response must include process-integrity validation, not just IT cleanup."
-      }
-    },
-    "framework_coverage": {
-      "NIST-800-82r3-ICS": {
-        "covered": true,
-        "adequate": false,
-        "gap": "OT/ICS devices often cannot be patched on an IT cadence (availability and safety constraints, vendor re-certification), so the exposure window for a KEV-listed flaw is months-to-years; the standard treats patching as primary but for OT the load-bearing control is network segmentation and access restriction."
-      },
-      "IEC-62443-3-3-zones-conduits": {
-        "covered": true,
-        "adequate": false,
-        "gap": "Zones-and-conduits segmentation is required, but it is frequently unimplemented or bypassed (flat OT networks, internet-exposed HMIs/cameras/PLCs); the device itself is exploitable whenever it is reachable, and the standard does not force a compromised-device response plan."
-      },
-      "NIS2-Art21-OT": {
-        "covered": true,
-        "adequate": false,
-        "gap": "Treats OT/ICS as essential-function infrastructure but lacks a device-level compressed remediation expectation and does not address the reality that many OT devices are end-of-life or unpatchable and must be isolated rather than fixed."
-      },
-      "ISO-27001-2022-A.8.8": {
-        "covered": true,
-        "adequate": false,
-        "gap": "'Appropriate timescales' is undefined and ill-suited to OT, where patching may be impossible; compensating controls (segmentation, monitoring, restricted engineering-workstation access) carry the load and must be explicit."
-      }
-    },
-    "compliance_exposure_score": {
-      "percent_audit_passing_orgs_still_exposed": 82,
-      "basis": "OpenPLC ScadaBR is OT/ICS equipment that frequently cannot be patched on an IT cadence and is often reachable on flat or internet-exposed networks; audited organizations that rely on a patch SLA rather than enforced segmentation remain exposed for this KEV-listed, actively-exploited flaw.",
-      "theater_pattern": "patch_management"
-    },
-    "ai_discovered_zeroday": false,
-    "ai_discovery_source": "vendor_research",
     "ai_assist_factor": "none",
     "new_control_requirements": [
       {
-        "id": "NEW-CTRL-118",
-        "name": "OT-DEFAULT-CREDENTIAL-ELIMINATION",
-        "description": "ScadaBR is an HMI web front end and has to be governed like one. The packet puts the upload-and-execute path at view_edit.shtm, so anything that can reach that page can reach JSP execution on the machine operators use to drive the process — the compromise lands on the supervisory host, not on a back-office web server. Put the ScadaBR interface on a segmented management plane rather than a flat plant network or a directly reachable address, and eliminate vendor-default or blank administrative accounts on every instance, because the packet's product description gates the upload behind a ScadaBR login and an unrotated default credential turns that gate into a formality. The distinguishing test: from a network segment with no operational need to reach the HMI, attempt to load view_edit.shtm — reachability alone is the finding, since past that point a single valid or default login yields code execution on the process-control host.",
-        "evidence": "The packet's vector states: 'OpenPLC ScadaBR contains an unrestricted upload of file with dangerous type vulnerability that allows remote authenticated users to upload and execute arbitrary JSP files via view_edit.shtm.' cwe_refs is CWE-434, cvss 8.8, rwep_score 77, cisa_kev true with kev_date 2025-12-03, active_exploitation 'confirmed', poc_available true. The packet's own citing gaps include NIS2-Art21-network-security ('Security of network and information systems') and UK-CAF-B4 ('System security').",
+        "id": "NEW-CTRL-021",
+        "name": "TIER-3-DEPENDENCY-INVENTORY",
+        "description": "The packet locates the defect in how React decodes payloads sent to React Server Function endpoints, which means the vulnerable code is React's own and an application's exposure is set by the React build linked into it — not by any entry an installed-software inventory would hold. In most deployments React is not the dependency the team declares: the application declares the server-rendering layer it builds on, and React arrives underneath that, so a supply-chain inventory that records only direct dependencies returns a clean answer for an application that is fully exploitable. Bound to this CVE, the control means the dependency inventory resolves and records, for every deployed application build, the React version actually linked into the artifact serving traffic, and that resolved record — not the top-level manifest — is what the vulnerability and KEV process queries. The packet also records that a second identifier, CVE-2025-66478, has been rejected but is associated with this entry, which is a further reason the inventory must be able to answer by component and version rather than only by a single CVE id. Distinguishing test: pick a deployed build rather than a repository, extract the React version from the artifact that is actually answering requests, and compare it to the fixed release — an estate that has merged the framework dependency bump in source while a built and deployed artifact still carries a pre-fix React remains exploitable with a patch attestation reading clean, and that divergence between source and artifact is the specific failure this control exists to surface. Precondition and limits: an inventory identifies which builds must be rebuilt, it does not remediate any of them. The packet registers no live-patch path and a vendor patch that typically requires a service restart, so each identified build must be rebuilt on the fixed React and the serving process restarted; a fixed artifact published but not rolled out is still serving the vulnerable decoder, and counting it as remediated is the same error one layer down.",
+        "evidence": "Packet: CWE-94 remote code execution in Meta React Server Components; unauthenticated remote code execution by exploiting a flaw in how React decodes payloads sent to React Server Function endpoints. The packet notes CVE-2025-66478 has been rejected but is associated with CVE-2025-55182. cisa_kev true, kev_date 2025-12-05, active_exploitation confirmed, poc_available true, CVSS 9.8, RWEP 83. patch_available true; live_patch_available false, live_patch_notes record no live-patch tool registered and a vendor patch that typically requires service restart or system reboot per the KEV requiredAction.",
         "gap_closes": [
-          "NIS2-Art21-network-security",
-          "UK-CAF-B4"
+          "ISO-27001-2022-A.8.8",
+          "AU-Essential-8-Patch"
         ]
       },
       {
         "id": "NEW-CTRL-001",
         "name": "CISA-KEV-RESPONSE-SLA",
-        "description": "The KEV clock for this entry runs against an HMI, where 'restart the service' is a process-availability decision rather than a maintenance chore — which is exactly why an unmodified KEV SLA gets waived on OT assets and the exposure persists. From the 2025-12-03 listing, the SLA must pre-authorize a ScadaBR restart window, because the packet records no live-patch tool and a vendor patch that typically requires a service restart or reboot; there is no in-place option that leaves the HMI running. Where the restart cannot be scheduled inside the window, the required output is a declared compensating control — blocking reachability to view_edit.shtm from anything but the operator segment — carried as an open, time-bound item, not an entry marked remediated on schedule.",
-        "evidence": "The packet records cisa_kev true with kev_date 2025-12-03, active_exploitation 'confirmed', poc_available true, cvss 8.8 and rwep_score 77 for CWE-434 on OpenPLC ScadaBR, with the upload path at view_edit.shtm. patch_available is true, live_patch_available is false, and live_patch_notes reads: 'No live-patch tool registered for this entry at bulk-import time. Vendor patch typically requires service restart or system reboot per the KEV requiredAction.'",
-        "gap_closes": [
-          "ISO-27001-2022-A.8.8",
-          "NIST-800-53-SI-2"
-        ]
-      }
-    ]
-  },
-  "CVE-2025-48633": {
-    "name": "Android Framework Information Disclosure Vulnerability",
-    "lesson_date": "2026-05-29",
-    "attack_vector": {
-      "description": "an out-of-bounds read information-disclosure flaw (CWE-125) in the Android Framework, used by a local app as a primitive in a privilege-escalation chain (leaking memory to defeat ASLR for a follow-on exploit). CISA KEV-listed 2025-12-02 with confirmed in-the-wild exploitation; this class forms the local-escalation half of a mobile-spyware chain.",
-      "privileges_required": "low (a local app or the foothold from an initial-access primitive)",
-      "complexity": "low — KEV-listed, actively exploited; treat as weaponized",
-      "ai_factor": "No AI involvement documented in discovery or weaponization."
-    },
-    "defense_chain": {
-      "prevention": {
-        "what_would_have_worked": "Apply the Android Security Bulletin OTA update promptly; enforce update SLAs via MDM on managed fleets, deploy mobile-threat-defense, and enable hardened/locked-down configurations for high-risk users.",
-        "was_this_required": true,
-        "framework_requiring_it": "CISA BOD 22-01 (KEV remediation)",
-        "adequacy": "The OTA fix is definitive; the gap is OEM/carrier patch reach and managed fleets that defer mobile updates."
-      },
-      "detection": {
-        "what_would_have_worked": "Mobile-threat-defense telemetry for unprivileged-to-elevated transitions and ASLR-defeating memory disclosure; vendor threat notifications for targeted users.",
-        "was_this_required": false,
-        "framework_requiring_it": null,
-        "adequacy": "Backstops unpatched devices; mobile-spyware chains are stealthy and frequently zero-click."
-      },
-      "response": {
-        "what_would_have_worked": "Force the OTA update; for a confirmed targeted device, preserve forensic state, rotate credentials and tokens stored on the device, and consider device replacement — spyware can persist across reboots.",
-        "was_this_required": true,
-        "framework_requiring_it": "NIST 800-53 IR-4",
-        "adequacy": "Mandatory for a KEV-listed mobile RCE/LPE; the exposure is every device that processed attacker content pre-patch."
-      }
-    },
-    "framework_coverage": {
-      "NIST-800-53-SI-2": {
-        "covered": true,
-        "adequate": false,
-        "gap": "The 30-day flaw-remediation SLA is far longer than the observed exploitation window for a KEV-listed, actively-exploited mobile flaw; commercial-surveillance and spyware chains weaponize these within days, and patch reach depends on OEM/carrier OTA cadence well beyond the vendor's release date."
-      },
-      "ISO-27001-2022-A.8.8": {
-        "covered": true,
-        "adequate": false,
-        "gap": "'Appropriate timescales' is undefined; the standard reading is unsafe for an actively-exploited mobile OS flaw, and the OEM/carrier OTA chain means many devices receive the fix weeks-to-never after disclosure."
-      },
-      "AU-ISM-1546": {
-        "covered": true,
-        "adequate": false,
-        "gap": "Essential 8 patch-applications (operating systems) is the right tier, but the load-bearing controls for mobile are vendor OTA cadence (Android Security Bulletin / Samsung SMR), MDM-enforced update SLAs on managed fleets, mobile-threat-defense, and hardened/locked-down configurations for high-risk users — none of which the framework names explicitly."
-      }
-    },
-    "compliance_exposure_score": {
-      "percent_audit_passing_orgs_still_exposed": 71,
-      "basis": "Android update reach depends on OEM/carrier OTA cadence; audited organizations that do not enforce mobile update SLAs via MDM remain exposed for this KEV-listed, actively-exploited flaw long after the fix is published.",
-      "theater_pattern": "patch_management"
-    },
-    "ai_discovered_zeroday": false,
-    "ai_discovery_source": "vendor_research",
-    "ai_assist_factor": "none",
-    "new_control_requirements": [
-      {
-        "id": "NEW-CTRL-126",
-        "name": "MOBILE-OS-MINIMUM-PATCH-LEVEL-ENFORCEMENT",
-        "description": "The Android Framework out-of-bounds read is closed only by advancing the handset's security-patch level, and the packet records no live-patch path - the device has to take the update and reboot. The MDM/EMM must read each managed device's security-patch level and block or quarantine handsets below the level carrying this Framework fix from organizational data, not merely surface the stale level on a dashboard. The second half matters more than usual here: the leak is consumed by a locally running app to defeat ASLR for the follow-on exploit, so devices that cannot yet update must also be prevented from installing untrusted or side-loaded apps - denying the local app that harvests the leak is the only mitigation available while the OEM update is out of reach. It is also the control that survives triage: at CVSS 5.5 this reads as a low-priority information disclosure, while the packet scores it RWEP 77 on confirmed in-the-wild use. Distinguishing test: enroll a device pinned below the fixing patch level and confirm it is denied protected resources.",
-        "evidence": "Packet vector: Android Framework contains an unspecified vulnerability that allows for information disclosure. attack_vector records an out-of-bounds read (CWE-125) in the Android Framework used by a local app as a primitive in a privilege-escalation chain, leaking memory to defeat ASLR for a follow-on exploit, and forming the local-escalation half of a mobile-spyware chain. CISA KEV-listed 2025-12-02, active_exploitation confirmed, PoC available, CVSS 5.5, RWEP 77. patch_available true; live_patch_available false, with live_patch_notes stating no live-patch tool is registered and that the vendor patch typically requires service restart or system reboot per the KEV requiredAction.",
+        "description": "For this entry the compressed tier is justified by the packet rather than by the CVSS band alone: the path needs no credential and no user interaction, a public PoC exists, exploitation is confirmed in the wild, and the flaw sits on an endpoint the application exposes to untrusted callers by design. The requirement bound to this product: from the 2025-12-05 KEV listing, every application instance that serves React Server Function endpoints is rebuilt on the fixed React and rolled out, with completion measured by the version answering traffic rather than by a merged dependency bump or an approved change record. Because the packet registers no live-patch path and states the vendor patch typically requires a service restart or system reboot, the clock ends at the restart of the serving process — an instance holding a fixed artifact it has not yet restarted onto is still running the vulnerable decoder and must be counted as exposed. On the compensating-control half the honest answer is that the packet supports none for this path, and saying otherwise would be the worse outcome: the vulnerable decode happens on the application's own public request path, which has to keep answering unauthenticated requests for the application to function, so unlike a management interface there is no segment to restrict it to and no feature flag the packet identifies as removing the surface. If a request-filtering layer is proposed as an interim measure it must be demonstrated on a staging instance to reject the payload before the decode runs; a filter that has not been shown to see the decoded payload is not a compensating control and must not be recorded as one. Taking an affected instance out of service is the only interim state the packet's facts actually support, and the remediation remains the rebuilt, restarted application.",
+        "evidence": "Packet: unauthenticated remote code execution (CWE-94) via a flaw in how React decodes payloads sent to React Server Function endpoints in Meta React Server Components. cisa_kev true with kev_date 2025-12-05, active_exploitation confirmed, poc_available true, CVSS 9.8, RWEP 83. patch_available true, live_patch_available false; live_patch_notes: no live-patch tool registered for this entry at bulk-import time, and the vendor patch typically requires service restart or system reboot per the KEV requiredAction.",
         "gap_closes": [
-          "AU-Essential-8-Patch",
-          "ISO-27001-2022-A.8.8",
           "NIST-800-53-SI-2",
-          "NIS2-Art21-patch-management"
-        ]
-      }
-    ]
-  },
-  "CVE-2025-48572": {
-    "name": "Android Framework Privilege Escalation Vulnerability",
-    "lesson_date": "2026-05-29",
-    "attack_vector": {
-      "description": "a privilege-escalation flaw (CWE-269) in the Android Framework, exploited by a local app to escalate privileges on the device (the local-escalation step after an initial-access primitive). CISA KEV-listed 2025-12-02 with confirmed in-the-wild exploitation; this class forms the local-escalation half of a mobile-spyware chain.",
-      "privileges_required": "low (a local app or the foothold from an initial-access primitive)",
-      "complexity": "low — KEV-listed, actively exploited; treat as weaponized",
-      "ai_factor": "No AI involvement documented in discovery or weaponization."
-    },
-    "defense_chain": {
-      "prevention": {
-        "what_would_have_worked": "Apply the Android Security Bulletin OTA update promptly; enforce update SLAs via MDM on managed fleets, deploy mobile-threat-defense, and enable hardened/locked-down configurations for high-risk users.",
-        "was_this_required": true,
-        "framework_requiring_it": "CISA BOD 22-01 (KEV remediation)",
-        "adequacy": "The OTA fix is definitive; the gap is OEM/carrier patch reach and managed fleets that defer mobile updates."
-      },
-      "detection": {
-        "what_would_have_worked": "Mobile-threat-defense telemetry for unprivileged-to-elevated transitions and ASLR-defeating memory disclosure; vendor threat notifications for targeted users.",
-        "was_this_required": false,
-        "framework_requiring_it": null,
-        "adequacy": "Backstops unpatched devices; mobile-spyware chains are stealthy and frequently zero-click."
-      },
-      "response": {
-        "what_would_have_worked": "Force the OTA update; for a confirmed targeted device, preserve forensic state, rotate credentials and tokens stored on the device, and consider device replacement — spyware can persist across reboots.",
-        "was_this_required": true,
-        "framework_requiring_it": "NIST 800-53 IR-4",
-        "adequacy": "Mandatory for a KEV-listed mobile RCE/LPE; the exposure is every device that processed attacker content pre-patch."
-      }
-    },
-    "framework_coverage": {
-      "NIST-800-53-SI-2": {
-        "covered": true,
-        "adequate": false,
-        "gap": "The 30-day flaw-remediation SLA is far longer than the observed exploitation window for a KEV-listed, actively-exploited mobile flaw; commercial-surveillance and spyware chains weaponize these within days, and patch reach depends on OEM/carrier OTA cadence well beyond the vendor's release date."
-      },
-      "ISO-27001-2022-A.8.8": {
-        "covered": true,
-        "adequate": false,
-        "gap": "'Appropriate timescales' is undefined; the standard reading is unsafe for an actively-exploited mobile OS flaw, and the OEM/carrier OTA chain means many devices receive the fix weeks-to-never after disclosure."
-      },
-      "AU-ISM-1546": {
-        "covered": true,
-        "adequate": false,
-        "gap": "Essential 8 patch-applications (operating systems) is the right tier, but the load-bearing controls for mobile are vendor OTA cadence (Android Security Bulletin / Samsung SMR), MDM-enforced update SLAs on managed fleets, mobile-threat-defense, and hardened/locked-down configurations for high-risk users — none of which the framework names explicitly."
-      }
-    },
-    "compliance_exposure_score": {
-      "percent_audit_passing_orgs_still_exposed": 71,
-      "basis": "Android update reach depends on OEM/carrier OTA cadence; audited organizations that do not enforce mobile update SLAs via MDM remain exposed for this KEV-listed, actively-exploited flaw long after the fix is published.",
-      "theater_pattern": "patch_management"
-    },
-    "ai_discovered_zeroday": false,
-    "ai_discovery_source": "vendor_research",
-    "ai_assist_factor": "none",
-    "new_control_requirements": [
-      {
-        "id": "NEW-CTRL-056",
-        "name": "MOBILE-ENDPOINT-MDM-ENFORCED-KEV-SLA",
-        "description": "The packet calls the Android Framework flaw unspecified and places the trigger in a local app escalating privileges on the device, which leaves the operator no component to disable and no setting to change — the platform update carrying the fix is the entire remediation, so the only questions are how fast it lands and whether the user can decline it. For this CVE that means the managed-device policy pushes the update inside the clock that opened with the 2025-12-02 KEV listing, with user deferral disallowed and the restart enforced, because the packet records no live-patch tool and states the vendor patch typically requires a service restart or system reboot per the KEV requiredAction: a handset that installed the update and never rebooted is still running the vulnerable framework and must be counted as exposed. Precondition, and it is the one that breaks this control on Android specifically: the operator can force installation only of a build the device's OEM or carrier has actually released for that model, so the SLA has two failure states that need separating — handsets that could have taken the update and were allowed to defer, which policy enforcement fixes, and models for which no fixed build has shipped, which policy cannot fix and which fall to the access-condition control below or to a replacement schedule. Distinguishing test: from the management console, list handsets by model and reported security patch level and show, per model, whether a fixed build exists and how many days elapsed between its availability and the device's restart onto it — an attestation that reports policy assignment is measuring the console, not the fleet.",
-        "evidence": "Packet vector: 'Android Framework contains an unspecified vulnerability that allows for privilege escalation.' attack_vector: 'a privilege-escalation flaw (CWE-269) in the Android Framework, exploited by a local app to escalate privileges on the device (the local-escalation step after an initial-access primitive).' Fields: cisa_kev true, kev_date 2025-12-02, active_exploitation confirmed, CVSS 7.8, RWEP 77, poc_available true, ai_discovered false, patch_available true, live_patch_available false, live_patch_notes 'No live-patch tool registered for this entry at bulk-import time. Vendor patch typically requires service restart or system reboot per the KEV requiredAction.'",
-        "gap_closes": [
-          "AU-Essential-8-Patch",
-          "NIST-800-53-SI-2"
+          "NIS2-Art21-vulnerability-handling"
         ]
       },
       {
-        "id": "NEW-CTRL-126",
-        "name": "MOBILE-OS-MINIMUM-PATCH-LEVEL-ENFORCEMENT",
-        "description": "Because the packet records the vulnerability itself as unspecified and the trigger as a local app escalating privilege on the handset, the operator's only remaining levers on a device that cannot yet take the fixed build are what that device is allowed to reach and what code it is allowed to run. The fixed Android security patch level therefore has to function as an access condition — mail, VPN and document access denied to any enrolled handset below it — rather than as a column on a compliance dashboard, because a device below the fix that keeps its access is exactly the device this escalation is aimed at: the packet places this CVE as the local-escalation half of a mobile-spyware chain, so the payoff is the organizational data and credentials on the handset. Precondition: restricting installs to a managed catalogue and blocking side-loading raises the bar for getting the attacker's app onto the device, but it does not evict an app already installed and it does not cover one that arrived through the normal store channel — a handset suspected of already running the escalating app belongs on the incident path with credential rotation, not on the install-policy path. Distinguishing test: enrol a device pinned below the fixed patch level and confirm the conditional-access policy actually denies it the protected resources, instead of flagging it non-compliant while its mail keeps syncing. This is a holding measure for the window before the fixed build and its required reboot land, not a substitute for them.",
-        "evidence": "The packet's own description of the flaw is 'an unspecified vulnerability that allows for privilege escalation', with the exploitation path 'exploited by a local app to escalate privileges on the device' and the framing that 'this class forms the local-escalation half of a mobile-spyware chain'. patch_available true with live_patch_available false and live_patch_notes stating the vendor patch typically requires a service restart or system reboot per the KEV requiredAction — so the interval this control covers is real and reboot-gated. active_exploitation confirmed; cisa_kev true, kev_date 2025-12-02; poc_available true; RWEP 77; CVSS 7.8; CWE-269.",
+        "id": "NEW-CTRL-032",
+        "name": "PERIMETER-COMPROMISE-REBUILD-NOT-PATCH",
+        "description": "The asset this control governs for this CVE is the application instance itself, not an appliance: it is the component that terminates untrusted requests, and the packet's path needs no credential and no user interaction — only the ability to send a payload to a React Server Function endpoint the application exposes. With active exploitation confirmed, a public PoC available, and the KEV listing dating to 2025-12-05, an instance that was reachable by an untrusted network while serving a pre-fix React during that window must default to the assume-compromise path: redeploy from a known-good build rather than upgrading the running instance in place, and rotate every secret and credential that application process could read. The upgrade closes the decode path; it removes nothing the attacker already placed and nothing already read. This is precisely the requirement the cited anti-malware control cannot meet — exploitation is code execution inside the application's own process, reached through an endpoint that is supposed to answer, so no file lands for signature scanning to match and a clean endpoint-protection report says nothing about whether this path was used. Preconditions, stated because this control is routinely over-applied: the assume-compromise default belongs to instances that were network-reachable on a pre-fix build during the exposure window, and does not extend to instances that were never reachable or that were already on the fixed release. Redeployment is a remediation only if the image being deployed carries the fixed React — rebuilding from the same pre-fix artifact reinstates the vulnerability while producing a rebuild record that reads like remediation. And rebuilding does nothing about material already exfiltrated: credentials the process held remain valid until they are rotated, which is why rotation is part of the requirement rather than a follow-up.",
+        "evidence": "Packet: unauthenticated remote code execution (CWE-94) in Meta React Server Components via a flaw in how React decodes payloads sent to React Server Function endpoints; active_exploitation confirmed; poc_available true; cisa_kev true with kev_date 2025-12-05; CVSS 9.8, RWEP 83. patch_available true, live_patch_available false, with the vendor patch typically requiring service restart or system reboot per the KEV requiredAction. The entry's citing framework gaps include CIS Controls v8 10.1 (Deploy and Maintain Anti-Malware Software).",
         "gap_closes": [
-          "ISO-27001-2022-A.8.8",
-          "NIS2-Art21-vulnerability-management",
-          "UK-CAF-B4"
+          "CIS-Controls-v8-10.1"
         ]
       }
     ]
   },
-  "CVE-2021-26829": {
-    "name": "OpenPLC ScadaBR Cross-site Scripting Vulnerability",
+  "CVE-2021-26828": {
+    "name": "OpenPLC ScadaBR Unrestricted Upload of File with Dangerous Type Vulnerability",
     "lesson_date": "2026-05-29",
     "attack_vector": {
-      "description": "a cross-site scripting flaw (CWE-79) on the ScadaBR SCADA/HMI web interface, letting an attacker run script in an operator's authenticated session. CISA KEV-listed 2025-11-28 with confirmed in-the-wild exploitation.",
+      "description": "an unrestricted file-upload flaw (CWE-434) on the ScadaBR SCADA/HMI web interface, letting an unauthenticated attacker upload a file (e.g. a web shell) for code execution on the HMI server. CISA KEV-listed 2025-12-03 with confirmed in-the-wild exploitation.",
+      "privileges_required": "none (the device is reachable by an unauthenticated attacker; exposure is amplified when the OT zone is not segmented)",
+      "complexity": "low — KEV-listed, actively exploited; treat as weaponized",
+      "ai_factor": "No AI involvement documented in discovery or weaponization."
+    },
+    "defense_chain": {
+      "prevention": {
+        "what_would_have_worked": "Apply the vendor firmware/update where one exists; where the device cannot be patched, isolate it in a segmented OT zone (zones-and-conduits / Purdue model), block all IT and internet reachability, and restrict access to authenticated engineering workstations.",
+        "was_this_required": true,
+        "framework_requiring_it": "CISA BOD 22-01 (KEV remediation) + IEC 62443-3-3",
+        "adequacy": "Patching is often impossible on OT; segmentation and access restriction are the real controls, and a flat or internet-exposed OT network defeats them."
+      },
+      "detection": {
+        "what_would_have_worked": "OT-network monitoring for unauthorized connections to the ScadaBR HMI, unexpected configuration/logic changes, and access from outside the device's intended zone.",
+        "was_this_required": false,
+        "framework_requiring_it": null,
+        "adequacy": "Necessary because unpatched OT devices may stay exploitable indefinitely; behavioral detection is the backstop."
+      },
+      "response": {
+        "what_would_have_worked": "Isolate the device, validate and restore its configuration/logic from a known-good baseline, rotate any credentials/keys it held, and investigate for safety-impacting manipulation; engage OT/safety engineering before any change.",
+        "was_this_required": true,
+        "framework_requiring_it": "NIST 800-82r3 + NIST 800-53 IR-4",
+        "adequacy": "Mandatory; OT compromise can have physical/safety consequences, so response must include process-integrity validation, not just IT cleanup."
+      }
+    },
+    "framework_coverage": {
+      "NIST-800-82r3-ICS": {
+        "covered": true,
+        "adequate": false,
+        "gap": "OT/ICS devices often cannot be patched on an IT cadence (availability and safety constraints, vendor re-certification), so the exposure window for a KEV-listed flaw is months-to-years; the standard treats patching as primary but for OT the load-bearing control is network segmentation and access restriction."
+      },
+      "IEC-62443-3-3-zones-conduits": {
+        "covered": true,
+        "adequate": false,
+        "gap": "Zones-and-conduits segmentation is required, but it is frequently unimplemented or bypassed (flat OT networks, internet-exposed HMIs/cameras/PLCs); the device itself is exploitable whenever it is reachable, and the standard does not force a compromised-device response plan."
+      },
+      "NIS2-Art21-OT": {
+        "covered": true,
+        "adequate": false,
+        "gap": "Treats OT/ICS as essential-function infrastructure but lacks a device-level compressed remediation expectation and does not address the reality that many OT devices are end-of-life or unpatchable and must be isolated rather than fixed."
+      },
+      "ISO-27001-2022-A.8.8": {
+        "covered": true,
+        "adequate": false,
+        "gap": "'Appropriate timescales' is undefined and ill-suited to OT, where patching may be impossible; compensating controls (segmentation, monitoring, restricted engineering-workstation access) carry the load and must be explicit."
+      }
+    },
+    "compliance_exposure_score": {
+      "percent_audit_passing_orgs_still_exposed": 82,
+      "basis": "OpenPLC ScadaBR is OT/ICS equipment that frequently cannot be patched on an IT cadence and is often reachable on flat or internet-exposed networks; audited organizations that rely on a patch SLA rather than enforced segmentation remain exposed for this KEV-listed, actively-exploited flaw.",
+      "theater_pattern": "patch_management"
+    },
+    "ai_discovered_zeroday": false,
+    "ai_discovery_source": "vendor_research",
+    "ai_assist_factor": "none",
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-118",
+        "name": "OT-DEFAULT-CREDENTIAL-ELIMINATION",
+        "description": "ScadaBR is an HMI web front end and has to be governed like one. The packet puts the upload-and-execute path at view_edit.shtm, so anything that can reach that page can reach JSP execution on the machine operators use to drive the process — the compromise lands on the supervisory host, not on a back-office web server. Put the ScadaBR interface on a segmented management plane rather than a flat plant network or a directly reachable address, and eliminate vendor-default or blank administrative accounts on every instance, because the packet's product description gates the upload behind a ScadaBR login and an unrotated default credential turns that gate into a formality. The distinguishing test: from a network segment with no operational need to reach the HMI, attempt to load view_edit.shtm — reachability alone is the finding, since past that point a single valid or default login yields code execution on the process-control host.",
+        "evidence": "The packet's vector states: 'OpenPLC ScadaBR contains an unrestricted upload of file with dangerous type vulnerability that allows remote authenticated users to upload and execute arbitrary JSP files via view_edit.shtm.' cwe_refs is CWE-434, cvss 8.8, rwep_score 77, cisa_kev true with kev_date 2025-12-03, active_exploitation 'confirmed', poc_available true. The packet's own citing gaps include NIS2-Art21-network-security ('Security of network and information systems') and UK-CAF-B4 ('System security').",
+        "gap_closes": [
+          "NIS2-Art21-network-security",
+          "UK-CAF-B4"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-001",
+        "name": "CISA-KEV-RESPONSE-SLA",
+        "description": "The KEV clock for this entry runs against an HMI, where 'restart the service' is a process-availability decision rather than a maintenance chore — which is exactly why an unmodified KEV SLA gets waived on OT assets and the exposure persists. From the 2025-12-03 listing, the SLA must pre-authorize a ScadaBR restart window, because the packet records no live-patch tool and a vendor patch that typically requires a service restart or reboot; there is no in-place option that leaves the HMI running. Where the restart cannot be scheduled inside the window, the required output is a declared compensating control — blocking reachability to view_edit.shtm from anything but the operator segment — carried as an open, time-bound item, not an entry marked remediated on schedule.",
+        "evidence": "The packet records cisa_kev true with kev_date 2025-12-03, active_exploitation 'confirmed', poc_available true, cvss 8.8 and rwep_score 77 for CWE-434 on OpenPLC ScadaBR, with the upload path at view_edit.shtm. patch_available is true, live_patch_available is false, and live_patch_notes reads: 'No live-patch tool registered for this entry at bulk-import time. Vendor patch typically requires service restart or system reboot per the KEV requiredAction.'",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.8",
+          "NIST-800-53-SI-2"
+        ]
+      }
+    ]
+  },
+  "CVE-2025-48633": {
+    "name": "Android Framework Information Disclosure Vulnerability",
+    "lesson_date": "2026-05-29",
+    "attack_vector": {
+      "description": "an out-of-bounds read information-disclosure flaw (CWE-125) in the Android Framework, used by a local app as a primitive in a privilege-escalation chain (leaking memory to defeat ASLR for a follow-on exploit). CISA KEV-listed 2025-12-02 with confirmed in-the-wild exploitation; this class forms the local-escalation half of a mobile-spyware chain.",
+      "privileges_required": "low (a local app or the foothold from an initial-access primitive)",
+      "complexity": "low — KEV-listed, actively exploited; treat as weaponized",
+      "ai_factor": "No AI involvement documented in discovery or weaponization."
+    },
+    "defense_chain": {
+      "prevention": {
+        "what_would_have_worked": "Apply the Android Security Bulletin OTA update promptly; enforce update SLAs via MDM on managed fleets, deploy mobile-threat-defense, and enable hardened/locked-down configurations for high-risk users.",
+        "was_this_required": true,
+        "framework_requiring_it": "CISA BOD 22-01 (KEV remediation)",
+        "adequacy": "The OTA fix is definitive; the gap is OEM/carrier patch reach and managed fleets that defer mobile updates."
+      },
+      "detection": {
+        "what_would_have_worked": "Mobile-threat-defense telemetry for unprivileged-to-elevated transitions and ASLR-defeating memory disclosure; vendor threat notifications for targeted users.",
+        "was_this_required": false,
+        "framework_requiring_it": null,
+        "adequacy": "Backstops unpatched devices; mobile-spyware chains are stealthy and frequently zero-click."
+      },
+      "response": {
+        "what_would_have_worked": "Force the OTA update; for a confirmed targeted device, preserve forensic state, rotate credentials and tokens stored on the device, and consider device replacement — spyware can persist across reboots.",
+        "was_this_required": true,
+        "framework_requiring_it": "NIST 800-53 IR-4",
+        "adequacy": "Mandatory for a KEV-listed mobile RCE/LPE; the exposure is every device that processed attacker content pre-patch."
+      }
+    },
+    "framework_coverage": {
+      "NIST-800-53-SI-2": {
+        "covered": true,
+        "adequate": false,
+        "gap": "The 30-day flaw-remediation SLA is far longer than the observed exploitation window for a KEV-listed, actively-exploited mobile flaw; commercial-surveillance and spyware chains weaponize these within days, and patch reach depends on OEM/carrier OTA cadence well beyond the vendor's release date."
+      },
+      "ISO-27001-2022-A.8.8": {
+        "covered": true,
+        "adequate": false,
+        "gap": "'Appropriate timescales' is undefined; the standard reading is unsafe for an actively-exploited mobile OS flaw, and the OEM/carrier OTA chain means many devices receive the fix weeks-to-never after disclosure."
+      },
+      "AU-ISM-1546": {
+        "covered": true,
+        "adequate": false,
+        "gap": "Essential 8 patch-applications (operating systems) is the right tier, but the load-bearing controls for mobile are vendor OTA cadence (Android Security Bulletin / Samsung SMR), MDM-enforced update SLAs on managed fleets, mobile-threat-defense, and hardened/locked-down configurations for high-risk users — none of which the framework names explicitly."
+      }
+    },
+    "compliance_exposure_score": {
+      "percent_audit_passing_orgs_still_exposed": 71,
+      "basis": "Android update reach depends on OEM/carrier OTA cadence; audited organizations that do not enforce mobile update SLAs via MDM remain exposed for this KEV-listed, actively-exploited flaw long after the fix is published.",
+      "theater_pattern": "patch_management"
+    },
+    "ai_discovered_zeroday": false,
+    "ai_discovery_source": "vendor_research",
+    "ai_assist_factor": "none",
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-126",
+        "name": "MOBILE-OS-MINIMUM-PATCH-LEVEL-ENFORCEMENT",
+        "description": "The Android Framework out-of-bounds read is closed only by advancing the handset's security-patch level, and the packet records no live-patch path - the device has to take the update and reboot. The MDM/EMM must read each managed device's security-patch level and block or quarantine handsets below the level carrying this Framework fix from organizational data, not merely surface the stale level on a dashboard. The second half matters more than usual here: the leak is consumed by a locally running app to defeat ASLR for the follow-on exploit, so devices that cannot yet update must also be prevented from installing untrusted or side-loaded apps - denying the local app that harvests the leak is the only mitigation available while the OEM update is out of reach. It is also the control that survives triage: at CVSS 5.5 this reads as a low-priority information disclosure, while the packet scores it RWEP 77 on confirmed in-the-wild use. Distinguishing test: enroll a device pinned below the fixing patch level and confirm it is denied protected resources.",
+        "evidence": "Packet vector: Android Framework contains an unspecified vulnerability that allows for information disclosure. attack_vector records an out-of-bounds read (CWE-125) in the Android Framework used by a local app as a primitive in a privilege-escalation chain, leaking memory to defeat ASLR for a follow-on exploit, and forming the local-escalation half of a mobile-spyware chain. CISA KEV-listed 2025-12-02, active_exploitation confirmed, PoC available, CVSS 5.5, RWEP 77. patch_available true; live_patch_available false, with live_patch_notes stating no live-patch tool is registered and that the vendor patch typically requires service restart or system reboot per the KEV requiredAction.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "ISO-27001-2022-A.8.8",
+          "NIST-800-53-SI-2",
+          "NIS2-Art21-patch-management"
+        ]
+      }
+    ]
+  },
+  "CVE-2025-48572": {
+    "name": "Android Framework Privilege Escalation Vulnerability",
+    "lesson_date": "2026-05-29",
+    "attack_vector": {
+      "description": "a privilege-escalation flaw (CWE-269) in the Android Framework, exploited by a local app to escalate privileges on the device (the local-escalation step after an initial-access primitive). CISA KEV-listed 2025-12-02 with confirmed in-the-wild exploitation; this class forms the local-escalation half of a mobile-spyware chain.",
+      "privileges_required": "low (a local app or the foothold from an initial-access primitive)",
+      "complexity": "low — KEV-listed, actively exploited; treat as weaponized",
+      "ai_factor": "No AI involvement documented in discovery or weaponization."
+    },
+    "defense_chain": {
+      "prevention": {
+        "what_would_have_worked": "Apply the Android Security Bulletin OTA update promptly; enforce update SLAs via MDM on managed fleets, deploy mobile-threat-defense, and enable hardened/locked-down configurations for high-risk users.",
+        "was_this_required": true,
+        "framework_requiring_it": "CISA BOD 22-01 (KEV remediation)",
+        "adequacy": "The OTA fix is definitive; the gap is OEM/carrier patch reach and managed fleets that defer mobile updates."
+      },
+      "detection": {
+        "what_would_have_worked": "Mobile-threat-defense telemetry for unprivileged-to-elevated transitions and ASLR-defeating memory disclosure; vendor threat notifications for targeted users.",
+        "was_this_required": false,
+        "framework_requiring_it": null,
+        "adequacy": "Backstops unpatched devices; mobile-spyware chains are stealthy and frequently zero-click."
+      },
+      "response": {
+        "what_would_have_worked": "Force the OTA update; for a confirmed targeted device, preserve forensic state, rotate credentials and tokens stored on the device, and consider device replacement — spyware can persist across reboots.",
+        "was_this_required": true,
+        "framework_requiring_it": "NIST 800-53 IR-4",
+        "adequacy": "Mandatory for a KEV-listed mobile RCE/LPE; the exposure is every device that processed attacker content pre-patch."
+      }
+    },
+    "framework_coverage": {
+      "NIST-800-53-SI-2": {
+        "covered": true,
+        "adequate": false,
+        "gap": "The 30-day flaw-remediation SLA is far longer than the observed exploitation window for a KEV-listed, actively-exploited mobile flaw; commercial-surveillance and spyware chains weaponize these within days, and patch reach depends on OEM/carrier OTA cadence well beyond the vendor's release date."
+      },
+      "ISO-27001-2022-A.8.8": {
+        "covered": true,
+        "adequate": false,
+        "gap": "'Appropriate timescales' is undefined; the standard reading is unsafe for an actively-exploited mobile OS flaw, and the OEM/carrier OTA chain means many devices receive the fix weeks-to-never after disclosure."
+      },
+      "AU-ISM-1546": {
+        "covered": true,
+        "adequate": false,
+        "gap": "Essential 8 patch-applications (operating systems) is the right tier, but the load-bearing controls for mobile are vendor OTA cadence (Android Security Bulletin / Samsung SMR), MDM-enforced update SLAs on managed fleets, mobile-threat-defense, and hardened/locked-down configurations for high-risk users — none of which the framework names explicitly."
+      }
+    },
+    "compliance_exposure_score": {
+      "percent_audit_passing_orgs_still_exposed": 71,
+      "basis": "Android update reach depends on OEM/carrier OTA cadence; audited organizations that do not enforce mobile update SLAs via MDM remain exposed for this KEV-listed, actively-exploited flaw long after the fix is published.",
+      "theater_pattern": "patch_management"
+    },
+    "ai_discovered_zeroday": false,
+    "ai_discovery_source": "vendor_research",
+    "ai_assist_factor": "none",
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-056",
+        "name": "MOBILE-ENDPOINT-MDM-ENFORCED-KEV-SLA",
+        "description": "The packet calls the Android Framework flaw unspecified and places the trigger in a local app escalating privileges on the device, which leaves the operator no component to disable and no setting to change — the platform update carrying the fix is the entire remediation, so the only questions are how fast it lands and whether the user can decline it. For this CVE that means the managed-device policy pushes the update inside the clock that opened with the 2025-12-02 KEV listing, with user deferral disallowed and the restart enforced, because the packet records no live-patch tool and states the vendor patch typically requires a service restart or system reboot per the KEV requiredAction: a handset that installed the update and never rebooted is still running the vulnerable framework and must be counted as exposed. Precondition, and it is the one that breaks this control on Android specifically: the operator can force installation only of a build the device's OEM or carrier has actually released for that model, so the SLA has two failure states that need separating — handsets that could have taken the update and were allowed to defer, which policy enforcement fixes, and models for which no fixed build has shipped, which policy cannot fix and which fall to the access-condition control below or to a replacement schedule. Distinguishing test: from the management console, list handsets by model and reported security patch level and show, per model, whether a fixed build exists and how many days elapsed between its availability and the device's restart onto it — an attestation that reports policy assignment is measuring the console, not the fleet.",
+        "evidence": "Packet vector: 'Android Framework contains an unspecified vulnerability that allows for privilege escalation.' attack_vector: 'a privilege-escalation flaw (CWE-269) in the Android Framework, exploited by a local app to escalate privileges on the device (the local-escalation step after an initial-access primitive).' Fields: cisa_kev true, kev_date 2025-12-02, active_exploitation confirmed, CVSS 7.8, RWEP 77, poc_available true, ai_discovered false, patch_available true, live_patch_available false, live_patch_notes 'No live-patch tool registered for this entry at bulk-import time. Vendor patch typically requires service restart or system reboot per the KEV requiredAction.'",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "NIST-800-53-SI-2"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-126",
+        "name": "MOBILE-OS-MINIMUM-PATCH-LEVEL-ENFORCEMENT",
+        "description": "Because the packet records the vulnerability itself as unspecified and the trigger as a local app escalating privilege on the handset, the operator's only remaining levers on a device that cannot yet take the fixed build are what that device is allowed to reach and what code it is allowed to run. The fixed Android security patch level therefore has to function as an access condition — mail, VPN and document access denied to any enrolled handset below it — rather than as a column on a compliance dashboard, because a device below the fix that keeps its access is exactly the device this escalation is aimed at: the packet places this CVE as the local-escalation half of a mobile-spyware chain, so the payoff is the organizational data and credentials on the handset. Precondition: restricting installs to a managed catalogue and blocking side-loading raises the bar for getting the attacker's app onto the device, but it does not evict an app already installed and it does not cover one that arrived through the normal store channel — a handset suspected of already running the escalating app belongs on the incident path with credential rotation, not on the install-policy path. Distinguishing test: enrol a device pinned below the fixed patch level and confirm the conditional-access policy actually denies it the protected resources, instead of flagging it non-compliant while its mail keeps syncing. This is a holding measure for the window before the fixed build and its required reboot land, not a substitute for them.",
+        "evidence": "The packet's own description of the flaw is 'an unspecified vulnerability that allows for privilege escalation', with the exploitation path 'exploited by a local app to escalate privileges on the device' and the framing that 'this class forms the local-escalation half of a mobile-spyware chain'. patch_available true with live_patch_available false and live_patch_notes stating the vendor patch typically requires a service restart or system reboot per the KEV requiredAction — so the interval this control covers is real and reboot-gated. active_exploitation confirmed; cisa_kev true, kev_date 2025-12-02; poc_available true; RWEP 77; CVSS 7.8; CWE-269.",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.8",
+          "NIS2-Art21-vulnerability-management",
+          "UK-CAF-B4"
+        ]
+      }
+    ]
+  },
+  "CVE-2021-26829": {
+    "name": "OpenPLC ScadaBR Cross-site Scripting Vulnerability",
+    "lesson_date": "2026-05-29",
+    "attack_vector": {
+      "description": "a cross-site scripting flaw (CWE-79) on the ScadaBR SCADA/HMI web interface, letting an attacker run script in an operator's authenticated session. CISA KEV-listed 2025-11-28 with confirmed in-the-wild exploitation.",
       "privileges_required": "none (the device is reachable by an unauthenticated attacker; exposure is amplified when the OT zone is not segmented)",
       "complexity": "low — KEV-listed, actively exploited; treat as weaponized",
       "ai_factor": "No AI involvement documented in discovery or weaponization."
@@ -20321,7 +20398,30 @@
     },
     "ai_discovered_zeroday": false,
     "ai_discovery_source": "vendor_research",
-    "ai_assist_factor": "none"
+    "ai_assist_factor": "none",
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-032",
+        "name": "PERIMETER-COMPROMISE-REBUILD-NOT-PATCH",
+        "description": "DELMIA Apriso is the manufacturing-operations server the packet names, and the packet puts unauthenticated remote code execution on it with confirmed in-the-wild exploitation and a public PoC. Applying the Dassault update closes the CWE-94 injection path and says nothing about what ran on the server while it was open, which is why the disposition for any Apriso instance that was reachable from an untrusted network on a vulnerable build is compromised-until-shown-otherwise rather than patched-therefore-clear. Applied to this product: capture and diff the Apriso application directories and configuration against known-good rather than closing on a post-update version check; hunt for content the application serves or executes that no sanctioned deployment placed there, because a web shell dropped through this path is application content rather than product code and the vendor update leaves it running; and rotate what the instance held — its application secrets and service credentials, including the accounts it uses to reach the systems on either side of it, since an MES exists to broker between the business layer and the plant floor and those credentials are the pivot code execution hands over. Extend the review to the manufacturing-operations data the server governs, because an integrity change made there survives the patch, produces no malware artefact, and is the outcome that matters on an OT-adjacent asset. Precondition: this disposition applies to instances whose exposure during the vulnerable window cannot be excluded; an instance that provably never answered an untrusted caller on a vulnerable build takes the vendor update on the KEV clock instead. Rebuilding or cleaning is not relief from the update — the instance must end up on the fixed build, and the packet records no live-patch path and a fix that typically requires a service restart or system reboot, so it is not remediated until that restart is taken.",
+        "evidence": "Packet fields for this entry: attack_vector 'a code-injection flaw (CWE-94) enabling unauthenticated remote code execution on the manufacturing-operations server. CISA KEV-listed 2025-10-28 with confirmed in-the-wild exploitation'; vector 'Dassault Systemes DELMIA Apriso contains a code injection vulnerability that could allow an attacker to execute arbitrary code'; cwe_refs CWE-94; cisa_kev true, kev_date 2025-10-28, active_exploitation confirmed, poc_available true, cvss 9.8, rwep_score 77; patch_available true, live_patch_available false, live_patch_notes 'No live-patch tool registered for this entry at bulk-import time. Vendor patch typically requires service restart or system reboot per the KEV requiredAction.' NIST-800-53-SI-2 (Flaw Remediation) and NIS2-Art21-vulnerability-management (Vulnerability handling) are recorded as citing gaps. The entry's own defense_chain records the cleanup requirement: prevention 'Apply the Dassault DELMIA Apriso update; hunt for web shells and rotate service credentials. DELMIA Apriso sits in the manufacturing-operations layer, so treat compromise as OT-adjacent', with adequacy 'Patch is necessary but, for these RCE/auth-bypass flaws, insufficient alone - web shells and stolen credentials survive the patch and require explicit cleanup', and response 'review data and downstream systems reachable from the DELMIA Apriso; assume compromise of accounts and managed endpoints in its reach' with adequacy 'patch-in-place without key rotation / web-shell hunting leaves the attacker resident or able to re-authenticate'.",
+        "gap_closes": [
+          "NIST-800-53-SI-2",
+          "NIS2-Art21-vulnerability-management"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-001",
+        "name": "CISA-KEV-RESPONSE-SLA",
+        "description": "The clock for this entry opened on the 2025-10-28 KEV listing, and for this product the hard part is what the fix costs: the packet records a vendor patch, live_patch_available false, and live_patch_notes stating the patch typically requires a service restart or system reboot per the KEV requiredAction. A manufacturing-operations server's restarts are scheduled against the production plan rather than against a vulnerability clock, so the specific way this remediation goes wrong is an Apriso instance where the update is installed and the restart is deferred to the next planned line stoppage — that instance is still running the vulnerable code and has to be counted as exposed, not as patched. Completion is therefore measured per Apriso instance from the build actually running after restart, not from an approved or downloaded state in the deployment console. Where the restart genuinely cannot be taken inside the window, this control's own alternative applies and must be recorded as what it is: documented compensating controls carrying a dated action item, an interim state rather than an SLA exception that closes the finding. Precondition on that interim, because it is where it gets over-claimed: the packet establishes the flaw as reachable by an unauthenticated attacker, so restricting which segments can reach the Apriso application interface bounds who can deliver the injection and nothing more — anything inside a permitted segment still reaches the CWE-94 sink, and no account-side or credential-side measure applies at all, because no account is used to get there.",
+        "evidence": "Packet fields for this entry: cisa_kev true with kev_date 2025-10-28, active_exploitation confirmed, poc_available true, cvss 9.8, rwep_score 77; cwe_refs CWE-94; attack_vector 'a code-injection flaw (CWE-94) enabling unauthenticated remote code execution on the manufacturing-operations server'; patch_available true, live_patch_available false, live_patch_notes 'No live-patch tool registered for this entry at bulk-import time. Vendor patch typically requires service restart or system reboot per the KEV requiredAction.' AU-Essential-8-Patch (Patch operating systems), ISO-27001-2022-A.8.8 (Management of technical vulnerabilities) and UK-CAF-B4 (System security) are recorded as citing gaps against this entry.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "ISO-27001-2022-A.8.8",
+          "UK-CAF-B4"
+        ]
+      }
+    ]
   },
   "CVE-2025-6205": {
     "name": "Dassault Systèmes DELMIA Apriso Missing Authorization Vulnerability",
@@ -21391,7 +21491,37 @@
     },
     "ai_discovered_zeroday": false,
     "ai_discovery_source": "vendor_research",
-    "ai_assist_factor": "none"
+    "ai_assist_factor": "none",
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-129",
+        "name": "ENTERPRISE-ERP-MANAGEMENT-FUNCTION-AUTHENTICATION",
+        "description": "The packet puts this defect in how the SKYSEA Client View management console program processes authentication on its TCP connection, and the outcome it records is an unauthenticated attacker bypassing that authentication, reaching privileged functionality and executing code. That is a fronting authentication step failing open, which is exactly the pattern this control governs: each privileged operation the management console exposes must authorize its own caller rather than inheriting a verdict from the connection-level authentication that precedes it, and the console's surface must be positioned so an untrusted caller cannot present a connection to it at all. This is also why the identity-and-access and least-privilege controls cited as insufficient on this entry can pass their attestations while the path stays wide open — the attacker never holds a SKYSEA operator account, so per-account privilege scoping is never consulted and the account model an identity review examines is bypassed rather than abused. Precondition: the per-function authorization is a property of the vendor's fixed build; this control states what to verify, it does not implement it. The packet records no live-patch path and a vendor patch that typically requires a service restart or system reboot, so a console whose package is updated but whose service has not been restarted is still running the vulnerable authentication handling and is not yet remediated. Distinguishing test: from a host with no management role, connect to a staging console and invoke each privileged function without authenticating, confirming each is refused before it runs.",
+        "evidence": "Packet: CWE-287 improper authentication; the vector states SKYSEA Client View \"contains an improper authentication vulnerability that allows remote code execution via a flaw in processing authentication on the TCP connection with the management console program\", and attack_vector describes an unauthenticated attacker bypassing authentication to reach privileged functionality. The entry cites UK CAF B2 (identity and access control) and NIST 800-53 AC-6 (least privilege) among the insufficient controls, neither of which the unauthenticated path engages. CVSS 9.8, RWEP 77, poc_available true, active_exploitation confirmed, CISA KEV-listed 2025-10-14. patch_available true; live_patch_available false, with the vendor patch typically requiring a service restart or system reboot.",
+        "gap_closes": [
+          "UK-CAF-B2"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-128",
+        "name": "APPSERVER-REMOTING-PROTOCOL-RESTRICTION",
+        "description": "The packet places the flaw on the TCP connection with the management console program — a non-HTTP listener that answers and processes authentication before the caller is authenticated, which is the class this control governs and the class a perimeter firewall and web-tier hardening never touch. For this deployment the requirement is that the SKYSEA console listener accept connections only from hosts that legitimately speak that protocol — the management servers and the endpoints they administer — enforced by network ACL or host firewall, rather than left reachable from any internal segment on the assumption that client-management software is internal by nature. With confirmed exploitation, a public exploit and a KEV listing on 2025-10-14 for an identifier issued in 2016, the vendor update and the service restart or reboot the packet says it requires belong on the KEV clock rather than in the next asset-management maintenance window. Precondition, stated plainly because it is where this control is usually over-claimed: an ACL bounds who can open the connection, it does not repair the authentication handling. Any compromised workstation, contractor laptop or jump host inside the permitted segment still reaches the vulnerable path, and on an estate where the console must accept connections from every managed endpoint, the permitted segment is effectively the estate — which leaves the update as the only measure that closes anything.",
+        "evidence": "Packet: attack_vector describes an improper-authentication flaw (CWE-287) in the SKYSEA Client View management server letting an unauthenticated attacker bypass authentication and reach privileged functionality, with the vector locating the flaw in authentication processing on the TCP connection with the management console program. CISA KEV-listed 2025-10-14 with active_exploitation confirmed and poc_available true; the entry cites NIS2 Art. 21 security of network and information systems as an insufficient control. patch_available true, live_patch_available false, vendor patch typically requiring a service restart or system reboot.",
+        "gap_closes": [
+          "NIS2-Art21-network-security"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-037",
+        "name": "FLEET-COMPROMISE-IR-PLAYBOOK",
+        "description": "The management console administers the endpoints running Client View, so an attacker who reaches privileged functionality on it is inside the system that manages the fleet rather than on a single host. Exploitation is confirmed and a public exploit exists, and the identifier is a 2016 CVE that was only KEV-listed on 2025-10-14 — so an operator finding an unpatched console now cannot treat the listing date as the start of the exposure window; an install that never took the fix has carried this path for as long as it has been running, and the playbook has to scope the window that way rather than to the last few weeks. For this product the playbook covers rotation of the credentials the console holds and of any credential that authenticated through it, review of console configuration and of anything the console could distribute to the endpoints it administers during the window, quarantine criteria for endpoints managed by a suspect console, and validation of those endpoints' state against a source the console does not control. Precondition on the last step: it exists only if such a source exists. An estate whose sole record of endpoint state is the console's own database has nothing independent to compare against, and that has to be arranged before the incident, not during it. Applying the vendor update and taking the restart it requires closes the entry path; it does not remove access an attacker established through it beforehand, which is precisely why closing this on the patch record leaves a compromised console counted as remediated.",
+        "evidence": "Packet: active_exploitation confirmed and poc_available true on a flaw described as unauthenticated remote code execution via the management console program's authentication processing; CVE identifier issued in 2016, CISA KEV-listed 2025-10-14; RWEP 77 against CVSS 9.8. The entry cites NIST 800-53 SI-2 (flaw remediation) and ASD Essential Eight patching among the insufficient controls. patch_available true with live_patch_available false and a vendor patch that typically requires a service restart or system reboot.",
+        "gap_closes": [
+          "NIST-800-53-SI-2",
+          "AU-Essential-8-Patch"
+        ]
+      }
+    ]
   },
   "CVE-2021-43798": {
     "name": "Grafana Path Traversal Vulnerability",
@@ -22175,7 +22305,39 @@
     },
     "ai_discovered_zeroday": false,
     "ai_discovery_source": "vendor_research",
-    "ai_assist_factor": "none"
+    "ai_assist_factor": "none",
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-124",
+        "name": "FRAMEWORK-DEFAULT-SECRET-DETECTION",
+        "description": "The credential this flaw turns on is inside the product, not on the operator's estate: the packet describes a hardcoded backdoor authentication credential in ScreenOS that grants administrative SSH/Telnet access to anyone presenting the planted password, reached as a supply-chain-planted backdoor. No amount of operator-side credential hygiene reduces the exposure, because the attacker never uses a firewall administrator account the operator created — per-account privilege scoping is never consulted and the account model an identity-and-access attestation examines is bypassed rather than abused, which is precisely why the identity-and-access-control and least-privilege gaps are recorded against this entry. Applied to this product, the requirement is to inventory every ScreenOS device in service with the firmware version it is actually running, gate each of those devices on the vendor's fixed release rather than on any password, multi-factor or administrator-role policy, and re-run that check after any firmware rollback, RMA replacement, or restore from an older image — those are the operations that quietly put a backdoored build back on a device that had been remediated. Distinguishing test: produce, per device, the ScreenOS version in service and show it is at or above the vendor's fixed release; an attestation that every firewall administrator holds a unique account with a rotated password and a second factor passes cleanly while this backdoor stays fully usable, because the planted credential is not one of those accounts and cannot be rotated by the operator at all. Precondition: the vendor update is what removes the embedded credential — this control is the inventory and the gate that get every device to it, and it offers nothing to a device that has not yet taken the update and the restart the packet records as required. It also detects nothing that has already happened; a device that was reachable by an untrusted network during the exposure window needs compromised-device treatment, not a version number.",
+        "evidence": "Packet: Juniper ScreenOS improper authentication (CWE-287); attack_vector describes a hardcoded backdoor authentication credential letting anyone with the planted password gain administrative SSH/Telnet access to the firewall, characterized as a supply-chain-planted backdoor; vector states the flaw could allow unauthorized remote administrative access to the device. CISA KEV-listed 2025-10-02, active_exploitation confirmed, poc_available true, CVSS 9.1, RWEP 77. patch_available true; live_patch_available false, live_patch_notes record no live-patch tool and a vendor patch that typically requires service restart or system reboot per the KEV requiredAction. Citing gaps include UK-CAF-B2 (Identity and access control) and NIST-800-53-AC-6 (Least Privilege).",
+        "gap_closes": [
+          "UK-CAF-B2",
+          "NIST-800-53-AC-6"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-032",
+        "name": "PERIMETER-COMPROMISE-REBUILD-NOT-PATCH",
+        "description": "This device is the perimeter, and the packet's outcome is not code execution on it but administrative control of it: unauthorized remote administrative access over SSH/Telnet to a ScreenOS firewall, with exploitation confirmed in the wild and a public proof of concept. An attacker holding the device's administrator role can add administrator accounts, alter security policy, and change VPN and management configuration — and installing the fixed ScreenOS release removes none of that. The update deletes the planted credential; it does not delete what was done with it, and an added administrator account survives the upgrade and needs no backdoor afterwards. So for any ScreenOS unit that was reachable by an untrusted network during the exposure window, remediation defaults to exporting the configuration for analysis, rebuilding the device onto the fixed release from vendor media, and rotating every credential the device held or authenticated — administrator passwords, VPN pre-shared keys, RADIUS/TACACS+ shared secrets, SNMP community strings, and any key or certificate stored on it — rather than upgrading in place and closing the ticket. Restoring the saved configuration onto the rebuilt device re-imports whatever the attacker changed, so that configuration is a forensic artifact and a reference for reconstruction, not the restore source. Distinguishing test: for each unit remediated, show the rebuild record and the credential-rotation record alongside the version evidence; a patch-cadence attestation recording the fixed release installed on every firewall is complete and passing while an attacker-added administrator account is still on the box. Precondition: rebuild bounds the device and nothing beyond it. Access the attacker gained through the firewall into the networks behind it persists after the rebuild, and every credential that transited or terminated on the device is suspect until rotated — that is an incident-response scope question, not a patching one.",
+        "evidence": "Packet: unauthorized remote administrative access to a Juniper ScreenOS firewall via a planted password (CWE-287, administrative SSH/Telnet access); active_exploitation confirmed; poc_available true; CISA KEV-listed 2025-10-02; CVSS 9.1; RWEP 77. patch_available true with live_patch_available false and live_patch_notes recording a vendor patch that typically requires service restart or system reboot per the KEV requiredAction. Citing gaps include AU-Essential-8-Patch, ISO-27001-2022-A.8.8 and NIST-800-53-SI-2, all of which treat installing the fixed release as remediation.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "ISO-27001-2022-A.8.8",
+          "NIST-800-53-SI-2"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-031",
+        "name": "FIREWALL-EGRESS-TELEMETRY-SEPARATE-TRUST-ZONE",
+        "description": "What this exploit emits is a successful administrative login. The packet describes a planted password that grants administrative SSH/Telnet access, so in the device's own record the attempt is indistinguishable from an operator signing in correctly on the first try: no failed-authentication burst, no brute-force pattern, no lockout, and no weak or expired credential for a password audit to find. The signal that survives is the authentication event correlated against who was actually working — an administrative SSH or Telnet session on a ScreenOS device that no change record, on-call roster, or jump-host session accounts for, and any administrative session sourced from outside the management network. That signal only exists off the device, because a caller who reaches the administrative role also holds the device's logging: ScreenOS must forward authentication and session logs continuously to a collector in a separate trust zone — different management plane, different credentials, different authentication path — and an interruption in that forwarding must itself raise an event rather than be written off as a collector fault. Distinguishing test: from an address outside the management network, complete an administrative SSH login against a staging ScreenOS device, then clear the device's local logs, and confirm the collector still holds the authentication record and raised the unaccounted-session alert. Alerting on failed logins, brute-force thresholds, or exploit-tool signatures would miss an attempt that behaves exactly as the packet describes, because the credential is valid on first presentation and nothing fails. Precondition, and it decides whether any of this is worth authoring: the forwarding has to have been running before the attempt. A rule written after the 2025-10-02 KEV listing against telemetry nobody was collecting produces nothing, and detection neither prevents the access nor evicts an attacker already administering the device — it bounds exposure to alert-and-response time during the window before the fixed release and its restart land.",
+        "evidence": "Packet: administrative SSH/Telnet access to the firewall granted to anyone presenting the planted password (CWE-287 improper authentication); vector states the flaw could allow unauthorized remote administrative access to the device. active_exploitation confirmed; poc_available true; CISA KEV-listed 2025-10-02. patch_available true, live_patch_available false, with the vendor patch typically requiring service restart or system reboot per the KEV requiredAction. Citing gap NIS2-Art21-network-security (Security of network and information systems).",
+        "gap_closes": [
+          "NIS2-Art21-network-security"
+        ]
+      }
+    ]
   },
   "CVE-2025-21043": {
     "name": "Samsung Mobile Devices Out-of-Bounds Write Vulnerability (variant: CVE-2025-21043)",
@@ -26913,7 +27075,40 @@
     },
     "ai_discovered_zeroday": false,
     "ai_discovery_source": "vendor_research",
-    "ai_assist_factor": "none"
+    "ai_assist_factor": "none",
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-055",
+        "name": "SECURITY-TOOL-INTEGRITY-VERIFICATION",
+        "description": "The vulnerable asset in this entry is the security-monitoring platform itself: the packet places a deserialization-of-untrusted-data path on the Wazuh server API and describes it as unauthenticated remote code execution on the security-monitoring server. For this product the control means the Wazuh server is enumerated and scanned as privileged, patch-prioritized software on the same clock as any other pre-auth RCE, instead of appearing in the estate's records only as a control that produces evidence about other hosts. Every Wazuh server in the deployment has to be reached, not just the one the console points at — a multi-node deployment has more than one manager, and the API surface exists on each. Because the packet records no live-patch path and notes the vendor patch typically requires a service restart or system reboot, a server whose package has been updated but whose service has not been restarted still runs the vulnerable code and must be counted as exposed rather than as remediated. Distinguishing test: from a host with no operational need to administer Wazuh, send an unauthenticated request carrying crafted serialized content at the server API on a staging manager and confirm it is refused before anything is deserialized — an estate whose vulnerability-management attestation covers the agents Wazuh watches, but never scans or tests the manager itself, passes cleanly while this path stays open.",
+        "evidence": "Packet: CWE-502 deserialization of untrusted data; vector states Wazuh \"contains a deserialization of untrusted data vulnerability that allows for remote code execution on Wazuh servers\", and attack_vector places it on the Wazuh server API as unauthenticated remote code execution on the security-monitoring server. CVSS 9.8, RWEP 77, poc_available true, active_exploitation confirmed, CISA KEV-listed 2025-06-10. patch_available true; live_patch_available false, with live_patch_notes recording no live-patch tool and a vendor patch that typically requires a service restart or system reboot.",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.8",
+          "NIST-800-53-SI-2",
+          "AU-Essential-8-Patch",
+          "UK-CAF-C1"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-125",
+        "name": "SERVICE-INTERNAL-PROTOCOL-DESERIALIZATION-HARDENING",
+        "description": "The packet's framing — untrusted data deserialized on a path a caller reaches without authenticating — makes the Wazuh server's listening surface a trust boundary rather than the internal convenience most deployments treat it as. Bound to this product, the control means content arriving at the manager is not permitted to drive object reconstruction, every listener the manager exposes requires peer authentication rather than only the one the operator's console happens to use, and those listeners are bound to the management and agent segment instead of to every interface. State the preconditions plainly, because this is where the control gets over-claimed: refusing to rebuild arbitrary objects from request content is a property the vendor's fixed build establishes — this control says what to verify, it does not implement it — and until that build and its service restart land, restricting reachability only bounds who can send the payload. It closes nothing for a caller already inside the permitted segment, and it is simply unavailable for the agent-facing listeners a manager must keep open to do its job. The least-privilege control cited as insufficient on this entry is not closed by anything here either: the attacker never holds a Wazuh account, so per-account privilege scoping is never consulted. Distinguishing test: from a segment with no monitoring role, open a connection to each listener a staging manager exposes and confirm the peer is rejected before any request content is deserialized.",
+        "evidence": "Packet: CWE-502 on the Wazuh server API, described as enabling unauthenticated remote code execution; poc_available true and active_exploitation confirmed, CISA KEV-listed 2025-06-10, RWEP 77 against CVSS 9.8. The entry cites NIST 800-53 AC-6 (Least Privilege) among its framework gaps, which the packet's own unauthenticated attack path never engages. patch_available true, live_patch_available false with a vendor patch that typically requires a service restart or system reboot.",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.8"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-037",
+        "name": "FLEET-COMPROMISE-IR-PLAYBOOK",
+        "description": "Compromise of a Wazuh server is not an endpoint incident — it is the loss of the plane through which every other detection in the estate is reported. With exploitation confirmed and a public exploit available, an operator who reaches the fixed build still has to answer what happened before it, and for the exposure window the manager's own alerts, rule state and stored telemetry are attacker-influenceable and cannot serve as the evidence that nothing occurred. The playbook this product needs covers rotation of the credentials and agent-enrolment material the manager holds, review of rule and configuration changes and of anything the manager could push to the agents it administers during the window, quarantine criteria for agents that reported to a suspect manager, and comparison of the manager's stored telemetry against a copy kept somewhere the manager cannot write. Precondition on that last step: it exists only if such a copy exists. An estate whose only record of its own monitoring output lives on the compromised host has no baseline to compare against and no way to establish the window, and building one after the fact produces nothing. Patching does not substitute for any of this — the update removes the deserialization path, it does not remove what an attacker left on a manager exploited before it landed.",
+        "evidence": "Packet: active_exploitation confirmed with poc_available true on a flaw described as unauthenticated remote code execution on the Wazuh server, CISA KEV-listed 2025-06-10. The entry records NIS2 Art. 21 incident handling and UK CAF C1 security monitoring among the framework controls that are insufficient here. patch_available true with live_patch_available false, so remediation is the vendor update plus the service restart or reboot the packet says it typically requires — neither of which removes attacker artefacts from a server exploited beforehand.",
+        "gap_closes": [
+          "NIS2-Art21-incident-handling",
+          "UK-CAF-C1"
+        ]
+      }
+    ]
   },
   "CVE-2024-42009": {
     "name": "RoundCube Webmail Cross-Site Scripting Vulnerability",
@@ -27800,7 +27995,38 @@
     },
     "ai_discovered_zeroday": false,
     "ai_discovery_source": "vendor_research",
-    "ai_assist_factor": "none"
+    "ai_assist_factor": "none",
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-042",
+        "name": "INCOMPLETE-PATCH-SEQUENCE-SEVERITY-MULTIPLIER",
+        "description": "The packet labels this a patch-bypass variant: the same traversal primitive on the same MagicINFO 9 Server file-handling surface, reached again after a fix was already issued for it. For this product that changes what \"remediated\" is permitted to mean. The MagicINFO server stays on an elevated review tier after the new build rather than dropping back to the generic application-patching cadence, and whatever restriction on who can reach its file-handling endpoints was put in place during the exposure window is retained rather than retired on the strength of the second fix — a primitive that has been bypassed once is evidence the class is likely to be reachable again, not evidence it is closed. The population this specifically catches is the operator whose vulnerability register shows the earlier CVE closed and whose scanner therefore reports the server compliant, while the bypass variant is the thing under confirmed exploitation. Distinguishing test: re-run the traversal the earlier fix was supposed to stop, with encoding and path-normalization variants, against a staging MagicINFO server on the new build, and confirm each is refused before any write occurs — an attestation that records \"vendor patch applied\" without re-testing the primitive cannot tell a fix from a second incomplete one.",
+        "evidence": "Packet: CWE-22 characterised in attack_vector as \"a patch-bypass variant\", letting an unauthenticated attacker write or read files outside the intended directory for code execution; the vector states Samsung MagicINFO 9 Server \"contains a path traversal vulnerability that allows an attacker to write arbitrary file as system authority\". CISA KEV-listed 2025-05-22, active_exploitation confirmed, poc_available true, RWEP 77 against CVSS 7.5. patch_available true; live_patch_available false, with the vendor patch typically requiring a service restart or system reboot.",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.8",
+          "NIS2-Art21-patch-management",
+          "AU-Essential-8-Patch"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-134",
+        "name": "DEVICE-MANAGEMENT-GATEWAY-CONFIG-ENDPOINT-AUTH-INPUT-NEUTRALIZATION",
+        "description": "MagicINFO 9 Server is the management plane for the displays it drives, and the defect sits on the file-handling path of that plane: a caller-supplied path selects where a write lands, and the packet records that write executing with system authority for a caller who never authenticated. Bound to this product, the control means every MagicINFO endpoint that accepts or names a file authorizes its caller before the content is processed at all, and resolves the destination to a canonical absolute path and confirms it remains inside the intended directory before the privileged write runs — so the caller cannot choose the target file. Alongside that, no MagicINFO server is left answering from a segment with no operational need to reach it: the administration console and the display fleet need it; a general user VLAN and the public internet do not. Preconditions: the endpoint-side authorization and path normalization are properties the vendor's fixed build establishes — this control states what to verify, not what to implement — and until that build and the service restart it requires land, restricting reachability bounds who can send the request while leaving the path fully open to anything inside the permitted segment. Where displays must reach the server across untrusted or remote networks, that restriction is not available at all and the update is the only lever. Distinguishing test: from a segment with no signage-management role, send an unauthenticated request whose destination path resolves outside the intended directory and confirm it is refused before anything is written.",
+        "evidence": "Packet: CWE-22 path traversal on Samsung MagicINFO 9 Server; vector states it \"allows an attacker to write arbitrary file as system authority\" and attack_vector describes an unauthenticated attacker writing or reading files outside the intended directory for code execution. CISA KEV-listed 2025-05-22 with active_exploitation confirmed and poc_available true. patch_available true, live_patch_available false, with the packet noting the vendor patch typically requires a service restart or system reboot.",
+        "gap_closes": [
+          "UK-CAF-B4"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-078",
+        "name": "ENDPOINT-MGMT-DEPLOYMENT-CHANNEL-INTEGRITY",
+        "description": "The primitive the packet records is an arbitrary file write as system authority on the server that stores and distributes content to the display fleet, so the exposed asset is not only a web root — it is every file that server holds and every file it hands out. Treat the MagicINFO installation and the directories it serves content from as a privileged distribution channel: file-integrity-monitor them and alert on writes that do not correspond to a sanctioned administrator action or a vendor update. This is the requirement that still has value after the fixed build lands, because the update closes the write path and removes nothing already written through it — with exploitation confirmed and a public exploit available, a file placed before remediation survives both the upgrade and the restart, and neither signature-based anti-malware nor authentication logging surfaces a bespoke artefact written through a traversal that never authenticated. Preconditions: monitoring detects, it does not prevent, so it bounds the window to alert-and-response time rather than closing the path; and it is only meaningful against a known-good baseline of those directories captured before the write. A server first inventoried after the exposure window records the attacker's file as part of normal state, which is why a server that was reachable during that window needs its served content compared against a known-good copy rather than being closed out on the patch.",
+        "evidence": "Packet: the vector records a write of an arbitrary file \"as system authority\", with attack_vector describing the writer as unauthenticated; active_exploitation confirmed, poc_available true, CISA KEV-listed 2025-05-22. patch_available true and live_patch_available false, with the vendor patch typically requiring a service restart or system reboot — remediation the packet supports for the write path, and which says nothing about content already written.",
+        "gap_closes": [
+          "NIST-800-53-SI-2"
+        ]
+      }
+    ]
   },
   "CVE-2023-38950": {
     "name": "ZKTeco BioTime Path Traversal Vulnerability",
@@ -27860,7 +28086,30 @@
     },
     "ai_discovered_zeroday": false,
     "ai_discovery_source": "vendor_research",
-    "ai_assist_factor": "none"
+    "ai_assist_factor": "none",
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-134",
+        "name": "DEVICE-MANAGEMENT-GATEWAY-CONFIG-ENDPOINT-AUTH-INPUT-NEUTRALIZATION",
+        "description": "The iclock API is BioTime's device-facing management channel, and the packet places the defect there: a path traversal (CWE-22) that lets an unauthenticated caller supply a crafted payload and read arbitrary files on the server. Bound to this product, the control means the iclock API authorizes its caller before acting on a request at all, and resolves any caller-supplied path to its canonical absolute form and confirms the result is still under the intended directory before the file handle is opened, rather than filtering traversal sequences out of the request string. The read direction is the whole exposure here: the packet's outcome is arbitrary file read on the biometric time-attendance server, so anything the service can open is returned to a caller who never authenticated and against whom no account model is ever consulted — which is why an identity or access attestation on BioTime operator accounts is never even reached by this path. Distinguishing test: from a host with no terminal role, send the iclock API unauthenticated requests whose path parameter resolves outside the intended directory on a staging BioTime instance, and confirm each is refused before anything is opened; a vulnerability-management attestation that lists BioTime as inventoried and scanned reads clean while this endpoint answers. Precondition, and it binds harder here than on most appliances: the endpoint-side authorization and path canonicalization are properties the vendor update establishes — this control states what to verify, it does not implement it. Until that update and the service restart the packet records have landed, restricting which network segments can reach the iclock API bounds who can send the crafted payload but leaves the endpoint fully exploitable to anything inside the permitted segment, and that lever is largely unavailable in the normal deployment, because the iclock API exists precisely so that time-attendance terminals spread across sites can reach it — the segment that must reach it is broad by design.",
+        "evidence": "Packet: ZKTeco BioTime, path traversal (CWE-22) in the iclock API that allows an unauthenticated attacker to read arbitrary files via supplying a crafted payload; attack_vector describes an unauthenticated attacker reading arbitrary files on the biometric time-attendance server. CISA KEV-listed 2025-05-19, active_exploitation confirmed, poc_available true, CVSS 7.5, RWEP 77. patch_available true; live_patch_available false, with live_patch_notes recording no live-patch tool for this entry and a vendor patch that typically requires service restart or system reboot per the KEV requiredAction. Citing gaps include UK-CAF-B4 (System security) and NIS2-Art21-network-security.",
+        "gap_closes": [
+          "UK-CAF-B4",
+          "NIS2-Art21-network-security"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-001",
+        "name": "CISA-KEV-RESPONSE-SLA",
+        "description": "For this entry the KEV clock opened 2025-05-19 against a flaw with confirmed in-the-wild exploitation and a public proof of concept, so the standard flaw-remediation cadence the citing gaps rest on is longer than the window an unauthenticated file-read on an internet-reachable attendance server actually gets. Two things make the clock behave differently here than on a generic application patch. First, the packet registers no live-patch tool and states the vendor patch typically requires a service restart or system reboot per the KEV requiredAction, so a BioTime server on which the installer has run but the service has not restarted is still serving the vulnerable iclock handler; completion has to be measured per server by the build actually running and the date the service last restarted, not by the installer-run date in a patch console. Second, and this is the half no patch-timescale control asks for: the update stops further reads, it does not un-disclose a file that was already read. Because exploitation is confirmed and a working proof of concept is public, every secret the service could open on a server that was reachable during the exposure window — database and integration credentials held in configuration, service-account material, anything else within the process's read scope — is treated as disclosed and rotated, with the rotation tracked to completion as its own item rather than folded into the patch ticket. Distinguishing test: produce, per BioTime server, the running build, the service restart timestamp, and the rotation record for the credentials that server holds; a flaw-remediation SLA marked met on installer-run date with no restart evidence and no rotation is the specific way this entry gets closed while it stays exploitable. Precondition: this control is a clock and an evidence standard, not a mitigation — it changes when the fixed build arrives and how completion is proven, and it gives an operator nothing during the interval before the update and restart land.",
+        "evidence": "Packet: CISA KEV-listed 2025-05-19 with active_exploitation confirmed and poc_available true; RWEP 77 against CVSS 7.5. patch_available true, live_patch_available false, live_patch_notes: no live-patch tool registered for this entry at bulk-import time, vendor patch typically requires service restart or system reboot per the KEV requiredAction. Vector: unauthenticated arbitrary file read via the iclock API. Citing gaps include NIST-800-53-SI-2 (Flaw Remediation), ISO-27001-2022-A.8.8 (Management of technical vulnerabilities) and AU-ISM-1546 (Patch operating systems and applications).",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.8",
+          "NIST-800-53-SI-2",
+          "AU-ISM-1546"
+        ]
+      }
+    ]
   },
   "CVE-2024-27443": {
     "name": "Synacor Zimbra Collaboration Suite (ZCS) Cross-Site Scripting (XSS) Vulnerability",
@@ -28063,7 +28312,30 @@
     },
     "ai_discovered_zeroday": false,
     "ai_discovery_source": "vendor_research",
-    "ai_assist_factor": "none"
+    "ai_assist_factor": "none",
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-040",
+        "name": "OWA-PER-REQUEST-SIEM-INGESTION",
+        "description": "The packet describes an attack that produces no authentication event: an attacker sends an HTML email, the victim views it in MDaemon's WorldClient webmail, and the injected JavaScript runs inside the session the victim already authenticated to, the recorded outcome being theft of session credentials and access to the mailbox. Bound to this product, the control means WorldClient's per-request access logs — not only its login and authentication records — are forwarded off the mail server to a collector the mail server's own compromise cannot reach, with retention long enough to cover a discovery lag measured in months. The detection has to key on what this exploitation actually emits, and the packet documents it: a burst of mailbox reads or message-fetch requests inside a single session immediately following the render of one message, mailbox operations attributed to a session whose source address differs from the address that authenticated it, and forwarding or filter changes made through the webmail session rather than through an administrative login. Authentication-only logging shows one ordinary successful login for all of that, and endpoint anti-malware has nothing to match, because no file is written and no process is spawned — the script executes in the browser session of a user who did exactly what webmail users do, which is why a rule keyed on process or file artifacts would miss an attempt behaving precisely as the packet describes. Distinguishing test: render a benign test message carrying an inert script in a staging WorldClient session and confirm the collector received the message-render request and the subsequent session-scoped mailbox requests as separate attributable records; if the only artifact is the login, the estate cannot reconstruct what a stolen session did. Precondition: this telemetry must already be collected and shipped before the attempt, so it is a posture held ahead of disclosure rather than a rule authored after one, and it detects rather than prevents. It does not invalidate a session token already stolen — a mailbox appearing in these records needs its sessions terminated, its credentials rotated and its forwarding rules and filters reviewed, which the packet's confirmed in-the-wild exploitation and public PoC make warranted wherever an unpatched WorldClient was reachable.",
+        "evidence": "Packet vector: 'MDaemon Email Server contains a cross-site scripting (XSS) vulnerability that allows a remote attacker to load arbitrary JavaScript code via an HTML e-mail message.' Attack vector places it in 'the MDaemon webmail (WorldClient), letting an attacker run script in a victim's authenticated session when they view a crafted email — used to steal session credentials and access the mailbox.' KEV listed 2025-05-19, active_exploitation 'confirmed', poc_available true, CVSS 6.1, RWEP 77, patch_available true, live_patch_available false. NIS2-Art21-incident-handling is recorded as a citing gap on this entry.",
+        "gap_closes": [
+          "NIS2-Art21-incident-handling"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-001",
+        "name": "CISA-KEV-RESPONSE-SLA",
+        "description": "This entry is the case the control exists for, because the number a CVSS-driven programme triages on is the wrong one here: CVSS 6.1 routes an XSS into a medium queue with a 30- or 90-day target, while the packet records CISA KEV listing on 2025-05-19, confirmed in-the-wild exploitation, a public PoC and RWEP 77. Applied to MDaemon, the clock runs from that KEV listing to the fixed WorldClient actually serving mail, and the measurement point is the one this product makes easy to miss: the packet registers no live-patch tool and states the vendor patch typically requires a service restart or system reboot per the KEV requiredAction, so a server where the update has been installed but the MDaemon service has not been restarted is still serving the vulnerable webmail and must be counted as exposed. On a mail server that restart is the step most likely to be pushed into a maintenance window, because taking it interrupts delivery and drops user sessions, and a deferral recorded as 'patched' is the specific way this remediation goes wrong. Distinguishing test: for each mail server, produce the installed version alongside the service start time and show the restart followed the update — a report clean on installed version alone cannot distinguish a remediated server from one still running vulnerable code from memory. Precondition: the packet offers no compensating control to stand in during the interval. The trigger is a user viewing a crafted message in the webmail, so the only interim levers reduce who can do that — restricting WorldClient reachability, or routing users to a mail client other than the vulnerable webmail — and neither removes the flaw; both are holding measures until the restart lands, and neither helps a session whose credentials were already stolen.",
+        "evidence": "Packet: CVSS 6.1 against RWEP 77, with CISA KEV listing 2025-05-19, active_exploitation 'confirmed' and poc_available true. patch_available true; live_patch_available false; live_patch_notes: 'No live-patch tool registered for this entry at bulk-import time. Vendor patch typically requires service restart or system reboot per the KEV requiredAction.' The exploitation path the packet gives is a victim viewing a crafted HTML email in the MDaemon webmail (WorldClient). Citing gaps on this entry include AU-Essential-8-Patch, NIST-800-53-SI-2 (Flaw Remediation), ISO-27001-2022-A.8.8 and UK-CAF-B4.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "NIST-800-53-SI-2",
+          "ISO-27001-2022-A.8.8",
+          "UK-CAF-B4"
+        ]
+      }
+    ]
   },
   "CVE-2025-4428": {
     "name": "Ivanti Endpoint Manager Mobile (EPMM) Code Injection Vulnerability (variant: CVE-2025-4428)",
@@ -28277,7 +28549,39 @@
     },
     "ai_discovered_zeroday": false,
     "ai_discovery_source": "vendor_research",
-    "ai_assist_factor": "none"
+    "ai_assist_factor": "none",
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-125",
+        "name": "SERVICE-INTERNAL-PROTOCOL-DESERIALIZATION-HARDENING",
+        "description": "The packet puts the sink in SAP NetWeaver's Visual Composer Metadata Uploader — a design-time upload function that answers on the same application-server surface as the business functions users must keep reaching. For this product the control means treating that upload path as a trust boundary rather than an internal convenience: content arriving there is constrained to the concrete types the metadata flow legitimately carries instead of being permitted to construct arbitrary objects, and the path answers only from segments with a design-time role rather than inheriting safety from the assumption that NetWeaver is an internal system. The packet carries two readings of who can reach it — its vector describes a privileged attacker deserializing untrusted or malicious content, its attack_vector records unauthenticated remote code execution on the application server — and the control has to hold under the broader one, which is exactly why the constraint belongs at the endpoint rather than in the surrounding session. This is also why the least-privilege gap recorded on this entry stays open under either reading: under the unauthenticated reading no account-privilege scoping is ever consulted, so an access-control attestation passes cleanly while the path is open; and under the privileged-caller reading the outcome the packet itself states is compromise of the confidentiality, integrity and availability of the host system, so SAP role scoping bounds what a user may do inside NetWeaver, not what a deserialization sink does to the host underneath it. Distinguishing test: against a staging instance, submit to the Visual Composer metadata-upload path an object of a type the metadata flow never legitimately conveys and confirm it is refused before any object is constructed. Precondition: the endpoint-side type constraint is a property the SAP update establishes — this control states what to verify, it does not implement it — and the reachability half holds only where the Visual Composer surface can be separated from the application-server surface that must stay reachable. Where they share one exposed surface, segmentation buys nothing and the update is the only remediation; the packet records that update as requiring a service restart or system reboot with no live-patch path, so an instance updated but not restarted is not yet remediated.",
+        "evidence": "Packet fields for this entry: vector 'SAP NetWeaver Visual Composer Metadata Uploader contains a deserialization vulnerability that allows a privileged attacker to compromise the confidentiality, integrity, and availability of the host system by deserializing untrusted or malicious content'; attack_vector 'a deserialization-of-untrusted-data flaw (CWE-502) on SAP NetWeaver (Visual Composer), enabling unauthenticated remote code execution on the application server. CISA KEV-listed 2025-05-15 with confirmed in-the-wild exploitation'; cwe_refs CWE-502; cisa_kev true, kev_date 2025-05-15, active_exploitation confirmed, poc_available true, cvss 9.8, rwep_score 77; patch_available true, live_patch_available false, live_patch_notes 'No live-patch tool registered for this entry at bulk-import time. Vendor patch typically requires service restart or system reboot per the KEV requiredAction.' NIST-800-53-AC-6 (Least Privilege), UK-CAF-B4 (System security) and NIS2-Art21-network-security (Security of network and information systems) are recorded as citing gaps against this entry.",
+        "gap_closes": [
+          "NIST-800-53-AC-6",
+          "UK-CAF-B4",
+          "NIS2-Art21-network-security"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-032",
+        "name": "PERIMETER-COMPROMISE-REBUILD-NOT-PATCH",
+        "description": "The packet records confirmed in-the-wild exploitation and a public PoC against a path to code execution on the SAP NetWeaver application server, and the entry's own defense chain states the consequence directly: web shells and stolen credentials survive the patch, and patch-in-place without cleanup leaves the attacker resident or with a usable internal-access pivot. So the disposition for any NetWeaver instance that was reachable from an untrusted network on a vulnerable build is compromised-until-shown-otherwise, not patched-therefore-clear. Applied to this product: hunt in the directories the Java stack serves and executes content from for artefacts that arrived outside a transport, because a web shell dropped through this path is application content rather than product code and the SAP update leaves it in place and running; rotate what the instance held — its application service accounts, the credentials it uses to reach its databases and the adjacent SAP systems, and any keys stored on the host — because NetWeaver is the business-critical hinge of the ERP estate and those credentials are the pivot that a host-level compromise hands over; and review for lateral movement out of the instance across the exposure window. The behaviour to hunt for is what this chain produces at execution time — the NetWeaver Java process spawning a shell or interpreter child, and files appearing in served directories with no corresponding transport — rather than a generic malware signature, which does not reliably fire on an in-process deserialization chain. Precondition: this disposition applies to instances whose exposure during the vulnerable window cannot be excluded; an instance that provably never answered an untrusted caller on a vulnerable build takes the vendor update on the KEV clock instead. Cleaning or rebuilding is not relief from the update — the instance must come up on the fixed level, and the packet records no live-patch path and a fix that typically requires a service restart or system reboot, so remediation is not complete until that restart is taken.",
+        "evidence": "Packet fields for this entry: attack_vector 'a deserialization-of-untrusted-data flaw (CWE-502) on SAP NetWeaver (Visual Composer), enabling unauthenticated remote code execution on the application server'; cisa_kev true, kev_date 2025-05-15, active_exploitation confirmed, poc_available true, cvss 9.8, rwep_score 77; patch_available true, live_patch_available false, live_patch_notes 'No live-patch tool registered for this entry at bulk-import time. Vendor patch typically requires service restart or system reboot per the KEV requiredAction.' NIST-800-53-SI-2 (Flaw Remediation) is recorded as a citing gap. The entry's own defense_chain supplies the cleanup requirement: prevention 'Apply the SAP NetWeaver update; hunt for web shells and rotate credentials - NetWeaver is business-critical and a compromise pivots into the ERP estate', adequacy 'Patch is necessary but insufficient alone - web shells and stolen credentials survive the patch'; detection 'Monitoring on the SAP NetWeaver: exploit-shaped requests, new web-shell files and unexpected child-process execution'; response 'Patch immediately, hunt and remove web shells, rotate application secrets/credentials, and review for lateral movement', adequacy 'patch-in-place without cleanup leaves the attacker resident or with a usable internal-access pivot'.",
+        "gap_closes": [
+          "NIST-800-53-SI-2"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-001",
+        "name": "CISA-KEV-RESPONSE-SLA",
+        "description": "The clock for this entry opened on the 2025-05-15 KEV listing, and the specific way an SAP estate misses it is the restart: the packet records a vendor patch, live_patch_available false, and live_patch_notes stating the fix typically requires a service restart or system reboot per the KEV requiredAction, while a production NetWeaver system's restarts are negotiated against business-processing windows rather than against a vulnerability clock. An instance where the SAP fix has been imported but the affected server has not been restarted is still running the vulnerable Visual Composer code and must be counted as exposed — that deferral, recorded as patched, is the failure mode this control exists to prevent, and it is more likely here than on a system whose owners can absorb an unplanned bounce. Completion is measured per NetWeaver instance running Visual Composer, from the level the instance actually reports after restart rather than from an import queue's status. Where an instance cannot take the restart inside the window, this control's alternative is documented compensating controls with a dated action item — for this CVE, restricting which callers can reach the Visual Composer metadata-upload surface — recorded as an interim state and not as an exception that closes the finding. Precondition on that interim: reachability restriction bounds who can deliver the payload and repairs nothing, and under the packet's own vector wording an attacker holding a privileged account, or one already positioned inside a permitted segment, still reaches the deserialization sink. Priority follows the packet rather than the CVSS band alone — confirmed in-the-wild exploitation and a public PoC against the server that carries the organisation's business transactions.",
+        "evidence": "Packet fields for this entry: cisa_kev true with kev_date 2025-05-15, active_exploitation confirmed, poc_available true, cvss 9.8, rwep_score 77; cwe_refs CWE-502; vector 'SAP NetWeaver Visual Composer Metadata Uploader contains a deserialization vulnerability that allows a privileged attacker to compromise the confidentiality, integrity, and availability of the host system by deserializing untrusted or malicious content'; patch_available true, live_patch_available false, live_patch_notes 'No live-patch tool registered for this entry at bulk-import time. Vendor patch typically requires service restart or system reboot per the KEV requiredAction.' AU-Essential-8-Patch (Patch operating systems) and ISO-27001-2022-A.8.8 (Management of technical vulnerabilities) are recorded as citing gaps against this entry.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "ISO-27001-2022-A.8.8"
+        ]
+      }
+    ]
   },
   "CVE-2024-12987": {
     "name": "DrayTek Vigor Routers OS Command Injection Vulnerability",
@@ -30192,7 +30496,28 @@
     },
     "ai_discovered_zeroday": false,
     "ai_discovery_source": "vendor_research",
-    "ai_assist_factor": "none"
+    "ai_assist_factor": "none",
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-122",
+        "name": "EOL-ASSET-DECOMMISSION",
+        "description": "This entry is the case the control exists for, and the packet states both halves of it: a vendor fix is recorded — MS10-002, dated 2010, requiring a reboot — and the 2026-05-20 KEV re-listing exists because long-tail unpatched and end-of-life estates remain exposed. Those facts together define the requirement: a host still exposed to this in 2026 is a host with no functioning update path, so reaching MS10-002 is an interim state and the terminal state is taking Internet Explorer out of service on that host, or replacing the host where the browser cannot be removed because the workload depends on it. Scope it to what the packet names — Internet Explorer's HTML rendering — and inventory Internet Explorer installs; the packet provides no mapping from this use-after-free into other browsers or other HTML-rendering software, so treating every renderer in the estate as an instance of this CVE manufactures findings and removal work against software no evidence implicates. Requirement in operational terms: enumerate every host that can still use Internet Explorer to render untrusted web content, record for each whether it can take MS10-002 and be rebooted, put those on the interim clock, and put every host that cannot on a dated removal or replacement schedule — a risk acceptance with no removal date leaves a KEV-listed flaw with a public PoC and confirmed exploitation in service indefinitely. Precondition on the interim half, which is where this is usually over-claimed: the packet's live-patch notes state MS10-002 requires a reboot, so a host that received the update but was not rebooted still runs the vulnerable code and is not remediated; and applying a 2010 patch closes this defect only, saying nothing about anything found in that browser since, which is exactly why the requirement cannot end at 'every install reports MS10-002'. Because exploitation is confirmed and the delivery path is a crafted page the victim visits, a host that browsed untrusted content while exposed belongs on the incident path rather than being closed on the patch record.",
+        "evidence": "Packet: CWE-416 use-after-free in Internet Explorer's HTML rendering allowing remote code execution when the victim visits a crafted page — the technique used in the 2010 Operation Aurora campaign. cisa_kev true, kev_date 2026-05-20, active_exploitation confirmed, poc_available true, CVSS 8.8, RWEP 70. The attack_vector states the legacy re-listing exists because long-tail unpatched/end-of-life estates remain exposed. patch_available true, live_patch_available false, live_patch_notes: 'Microsoft patch MS10-002 (2010); requires reboot.'",
+        "gap_closes": [
+          "NIS2-Art21-patch-management",
+          "NIST-800-53-SI-2"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-018",
+        "name": "SCANNER-PAPER-COMPLIANCE-TEST",
+        "description": "The paper-compliance outcome on this entry is a fleet report showing full MS10-002 coverage, because the population the scanner can see is the managed, modern one — and the packet places the exposure in long-tail unpatched and end-of-life estates, which are precisely the hosts no management agent enrols and no credentialed scan authenticates to. A patch-coverage percentage computed over enrolled hosts is silent about unenrolled ones, and the metric improves as the exposed hosts drop out of the denominator, so the number gets better as the risk gets worse. The operational test has two halves. First, reconcile the patch-management inventory against an independent observation of what is actually on the network — DHCP leases, switch MAC tables, or network discovery — and show that every host observed there is either enrolled and reporting Internet Explorer's state, or on a dated removal schedule; an unreconciled coverage figure is the paper-compliance answer and must not be accepted as evidence. Second, for hosts that do report, confirm the reboot has actually happened, because the packet's remediation note states MS10-002 requires a reboot while most consoles mark a host patched on installation — an installed-not-rebooted host counts as compliant while still running the vulnerable code. Keep the test population to hosts running Internet Explorer as the packet names it, rather than every HTML renderer in the estate. Limits: this control makes the verdict truthful, it does not change the exposure. Hosts the reconciliation surfaces still need MS10-002 and a reboot if they can take it, and a dated removal schedule if they cannot.",
+        "evidence": "Packet: the attack_vector states the legacy re-listing exists because long-tail unpatched/end-of-life estates remain exposed; cisa_kev true with kev_date 2026-05-20 and active_exploitation confirmed; poc_available true; CWE-416 use-after-free in Internet Explorer's HTML rendering triggered when the victim visits a crafted page; CVSS 8.8, RWEP 70. patch_available true, live_patch_available false, live_patch_notes: 'Microsoft patch MS10-002 (2010); requires reboot.'",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.8"
+        ]
+      }
+    ]
   },
   "CVE-2010-0806": {
     "name": "Microsoft Internet Explorer Use-After-Free (iepeers)",
@@ -30247,7 +30572,29 @@
     },
     "ai_discovered_zeroday": false,
     "ai_discovery_source": "vendor_research",
-    "ai_assist_factor": "none"
+    "ai_assist_factor": "none",
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-122",
+        "name": "EOL-ASSET-DECOMMISSION",
+        "description": "A vendor fix for this specific CVE exists — the packet records Microsoft patch MS10-018 (2010), requiring a reboot — but the browser carrying it has no ongoing patch path: the entry's own framework_coverage states that Internet Explorer is end-of-life and unsupported, and the packet's attack_vector explains the 2026-05-20 re-listing as existing because long-tail unpatched and end-of-life estates remain exposed. So MS10-018 plus its reboot is the interim state on any host that can still take it, and removing Internet Explorer is the terminal one. Applied here: enumerate every host on which Internet Explorer can still render attacker-supplied web content, including hosts where it survives not as the user's browser but as the renderer behind a legacy line-of-business application, and carry each to removal or replacement on a bounded, dated schedule rather than an open-ended risk acceptance. Scope it to Internet Explorer as the packet does — the packet ties the use-after-free to that browser's iepeers component and gives no mapping into other browsers or HTML renderers, so putting every browser in the estate on a removal schedule manufactures work against software no evidence implicates. Precondition, and the reason the interim measures are a holding action rather than a fix: restricting Internet Explorer to a named allowlist of internal legacy applications narrows the paths by which an attacker-controlled page reaches the vulnerable renderer, but it covers only the paths the policy actually mediates and leaves the vulnerable component installed and reachable — an allowlisted internal application that loads third-party content, or a redirect out of one, still lands on the sink. Distinguishing test: on a representative host, produce evidence that Internet Explorer is gone, not evidence that it has been demoted from default browser or configured with hardened zones — a host that keeps the renderer while the policy points users elsewhere still carries the vulnerable code.",
+        "evidence": "Packet fields for this entry: vector 'A use-after-free in Internet Explorer's iepeers.dll allows remote code execution when the victim visits a crafted page'; attack_vector 'a use-after-free (CWE-416) in the Internet Explorer iepeers component, exploitable by an attacker-controlled web page for code execution in the browser. CISA KEV-listed 2026-05-20 with confirmed in-the-wild exploitation; the legacy re-listing exists because long-tail unpatched/end-of-life estates remain exposed'; cwe_refs CWE-416; cisa_kev true, kev_date 2026-05-20, active_exploitation confirmed, poc_available true, cvss 8.8, rwep_score 70; patch_available true, live_patch_available false, live_patch_notes 'Microsoft patch MS10-018 (2010); requires reboot.' AU-Essential-8-App-Hardening (User application hardening) and UK-CAF-B4 (System security) are recorded as citing gaps. The entry's own framework_coverage supplies the end-of-life fact: its ISO-27001-2022-A.8.8 gap states 'the legacy re-listing exists because organizations still run vulnerable browser/reader builds (Internet Explorer is end-of-life and unsupported)', its AU-ISM-1546 gap names 'retiring end-of-life software (Internet Explorer)' among the load-bearing controls, and its defense_chain prevention step reads 'retire end-of-life software such as Internet Explorer'.",
+        "gap_closes": [
+          "AU-Essential-8-App-Hardening",
+          "UK-CAF-B4"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-001",
+        "name": "CISA-KEV-RESPONSE-SLA",
+        "description": "The KEV listing on this entry is dated 2026-05-20, not 2010, and this control's clock runs from the KEV listing or patch availability, whichever is later — so a sixteen-year-old CVE carries a clock that opened this year. That has to be said explicitly because of how a vulnerability programme misses this one, which is structural rather than deliberate: a KEV-driven remediation queue that filters or de-prioritises pre-2015 CVE identifiers as historical, and a scanner whose browser coverage only knows how to grade supported builds, will both stay silent while CISA is asserting confirmed in-the-wild exploitation now. The verified mitigation the clock demands for this CVE is MS10-018 with the reboot the packet records, applied to every host that can still take it; for the hosts that cannot — which per the entry's own attack_vector is the population the re-listing exists for — it is a documented compensating control carrying a dated removal plan, not an open exception. Verification has to read the state of the browser component on the host itself rather than a management console's verdict: on a long-tail or end-of-life estate the console's approved or not-applicable status is the artefact most likely to be stale, and that staleness is precisely how these hosts survived sixteen years of patch cycles while the compliance record stayed clean.",
+        "evidence": "Packet fields for this entry: cve CVE-2010-0806 with cisa_kev true and kev_date 2026-05-20, active_exploitation confirmed, poc_available true, cvss 8.8, rwep_score 70; patch_available true, live_patch_available false, live_patch_notes 'Microsoft patch MS10-018 (2010); requires reboot.'; attack_vector records that 'the legacy re-listing exists because long-tail unpatched/end-of-life estates remain exposed'. ISO-27001-2022-A.8.8 (Management of technical vulnerabilities) and NIST-800-53-SI-2 (Flaw Remediation) are recorded as citing gaps against this entry.",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.8",
+          "NIST-800-53-SI-2"
+        ]
+      }
+    ]
   },
   "CVE-2009-1537": {
     "name": "Microsoft DirectShow QuickTime Parsing Memory Corruption",
@@ -36490,7 +36837,31 @@
     },
     "ai_discovered_zeroday": false,
     "ai_discovery_source": "human_researcher",
-    "ai_assist_factor": "none"
+    "ai_assist_factor": "none",
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-001",
+        "name": "CISA-KEV-RESPONSE-SLA",
+        "description": "The remediation-priority problem this entry poses is a timing one, and it is specific: the identifier is a 2023 CVE but the KEV listing is dated 2025-02-25, and the packet records no public proof-of-concept, which is what holds RWEP at 46 against a CVSS of 9.0. An estate that orders its Zimbra patch queue by CVE age, by disclosure date, or by whether an exploit is publicly circulating will place this behind newer and noisier items while the packet records exploitation as confirmed. The control's requirement here is that the KEV listing date, not the CVE year and not exploit-availability signal, starts the clock for every Zimbra Collaboration Suite instance in the 8.8.15 line the packet names, and that completion is measured by each instance actually running the vendor's fixed build rather than by a change ticket being approved. Scope it to the Zimbra Collaboration Suite instances the packet identifies; nothing here maps the flaw into other groupware in the estate. Preconditions. The packet records no live-patch path for this product class and names the vendor update as the remediation, so the only two states the SLA can be met in are 'fixed build running' or 'documented compensating controls in place with a dated action item' — there is no third state where the software keeps running unmodified and the requirement is satisfied. And a 4-hour clock is only real where an emergency change path exists for the mail platform that does not queue behind the monthly window; naming the SLA without pre-authorising that path produces a missed target rather than a fast one.",
+        "evidence": "Packet facts: name 'Synacor Zimbra Collaboration Suite (ZCS) Cross-Site Scripting (XSS) Vulnerability', CWE-79; CISA KEV-listed 2025-02-25 with active_exploitation 'confirmed'; CVSS 9.0, RWEP 46; poc_available false; patch_available true, live_patch_available false, with live_patch_notes stating 'No live-patch path for this product class; remediation is the vendor update ... plus the named compensating controls until it lands'. Vector: 'Cross Site Scripting vulnerability in Zimbra ZCS v.8.8.15 allows a remote authenticated attacker to execute arbitrary code via a crafted script to the /h/autoSaveDraft function.'",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "ISO-27001-2022-A.8.8",
+          "NIST-800-53-SI-2",
+          "NIS2-Art21-vulnerability-management",
+          "UK-CAF-B4"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-040",
+        "name": "OWA-PER-REQUEST-SIEM-INGESTION",
+        "description": "This control's substance is webmail per-request access logging shipped off the mail server, and the asset class it was written for is the same one this packet names — a web mail client whose exploitation stage rides an already-authenticated session. Applied to Zimbra Collaboration Suite, it means the web client's per-request access logs, not just its authentication events, are forwarded to an external collector with retention long enough to cover the window this CVE actually spent unremediated. The behaviour to key on is the one the packet documents: a request to the /h/autoSaveDraft function carrying script content, issued by an authenticated ZCS account, and the subsequent rendering of that draft in another user's session. Both halves matter — the packet's attacker is a remote authenticated attacker, so the request arrives inside a valid session and the authentication log records nothing but a normal login by an account that is entitled to log in. Nothing crashes, no privilege is requested, and no failed authentication occurs, so a detection built on login anomalies, lockouts, or availability loss misses an attempt that behaves exactly as described. This is also the only way an operator answers the question the KEV listing forces: the packet records confirmed in-the-wild exploitation with no public proof-of-concept, so the technique is not reproducible in-house and the sole evidence of whether a given instance was targeted is what its request logs preserved. Preconditions. The logging and forwarding must have been in place during the exposure window — configuring it now answers nothing about the past, and logs left on the mail server are within reach of anyone who leveraged this to act as a mailbox owner. And detection is not remediation: it bounds the response window and scopes an investigation, while the vendor update the packet names remains the fix.",
+        "evidence": "Packet facts: vector states the flaw 'allows a remote authenticated attacker to execute arbitrary code via a crafted script to the /h/autoSaveDraft function' in Zimbra ZCS v.8.8.15; CWE-79; active_exploitation 'confirmed' with CISA KEV listing 2025-02-25; poc_available false; patch_available true, live_patch_available false with the vendor update recorded as the remediation.",
+        "gap_closes": [
+          "NIS2-Art21-vulnerability-management"
+        ]
+      }
+    ]
   },
   "CVE-2024-49035": {
     "name": "Microsoft Partner Center Improper Access Control Vulnerability",
@@ -36917,7 +37288,29 @@
         "adequate": false,
         "gap": "Secure-by-default configuration for the CMS platform doesn't extend to an unauthenticated file-attachment feature bundled inside a calendaring extension."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-001",
+        "name": "CISA-KEV-RESPONSE-SLA",
+        "description": "iCagenda is a Joomla extension, and that is what makes a KEV clock the missing control here rather than a redundant one: the remediation this flaw needs is an update to the extension, which does not arrive with Joomla core updates and is not what a 'the CMS is current' attestation measures. Bound to this CVE the requirement is that the clock opened by the 2026-07-10 KEV listing drives the iCagenda update on every site in the estate that has the extension installed, with completion measured as the installed iCagenda version per site rather than as a ticket closed against the platform. Because the packet's path is a single unauthenticated request to a public event-attachment feature that lands an executing PHP file, there is no authentication or authorization step for a compensating control to tighten — the only levers before the update lands are removing or disabling the attachment feature and denying script execution in the directory uploads land in. State the precondition on that interim measure rather than recording it as the mitigation: an execution-deny holds only where the operator controls the web-server configuration and where the attachment directory is the only place the upload can come to rest; it is unavailable on shared hosting where that configuration is not the operator's, and it does not remove a file already written. With exploitation confirmed and a public PoC, a site that ran the vulnerable extension while reachable may already hold an uploaded PHP file, and that file survives the update — so the update closes the path and the site still needs its attachment directory examined for content it did not put there.",
+        "evidence": "Packet: CWE-434, unrestricted upload of file with dangerous type; an unauthenticated attacker submits a crafted file through iCagenda's public event-attachment feature, the extension does not validate file type or require authentication, and a PHP payload lands in a web-accessible directory and executes. CISA KEV-listed 2026-07-10; active_exploitation confirmed; poc_available true; CVSS 9.8; RWEP 63. patch_available true, live_patch_available false, with no live-patch note recorded.",
+        "gap_closes": [
+          "NIST-800-53-SI-2",
+          "NIS2-Art21-vulnerability-handling"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-018",
+        "name": "SCANNER-PAPER-COMPLIANCE-TEST",
+        "description": "Paper compliance has a specific shape on this CVE: a scan that reports the Joomla core build says nothing about iCagenda, so a site with a fully patched CMS and a vulnerable extension reports clean, and the configuration-management record backing that report is complete at the platform level and empty at the extension level. The operational test for this entry is whether the scan enumerates installed Joomla extensions with their versions and compares iCagenda's installed version against the fixed one — and, because the packet's path needs no credentials and no authentication, whether an unauthenticated probe of the site's event-attachment endpoint is part of the check rather than a credentialed inventory the scanner may not be able to obtain. That second half is what distinguishes a real test from a version-table lookup: the exploit is an unauthenticated upload to a public feature, so the check that mirrors the attacker's position is an unauthenticated one. Precondition and limits: this control determines whether the estate can see the exposure, it does not remove it, and a site the scan flags still needs the extension update the packet records. It also depends on the extension inventory being obtainable at all — sites where extensions were installed outside the operator's change process are exactly the sites a credentialed scan under-reports, so a maintained record of which extensions each Joomla site runs is what makes the scan meaningful rather than reassuring.",
+        "evidence": "Packet: the affected component is the iCagenda extension for Joomla, not the Joomla platform; the flaw is CWE-434 and the path is an unauthenticated crafted file submitted through the extension's public event-attachment feature, landing a PHP payload in a web-accessible directory where it executes. CISA KEV-listed 2026-07-10; active_exploitation confirmed; poc_available true; CVSS 9.8; RWEP 63; patch_available true.",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.9",
+          "UK-CAF-B4"
+        ]
+      }
+    ]
   },
   "CVE-2026-56290": {
     "name": "Joomlack Page Builder Improper Access Control Vulnerability",
@@ -37197,7 +37590,39 @@
         "adequate": false,
         "gap": "Boundary protection guidance recommends restricting the management interface to trusted internal IPs, but this is advisory in PAN-OS deployments — it doesn't enforce network isolation by default, leaving internet-exposed interfaces reachable pre-patch."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-030",
+        "name": "PERIMETER-DEVICE-ZERODAY-SLA-TIER",
+        "description": "The vulnerable surface here is the PAN-OS management web interface, and the packet records the vendor's own risk-reduction step for it: restricting access to the management web interface to only trusted internal IP addresses. Bound to this deployment the control means two things. First, every PAN-OS firewall's management web interface answers only from a named management network — not from the internet and not from a general user VLAN — with that reachability enforced by the surrounding network rather than asserted from 'the management interface is internal'. Second, the remediation clock for the vendor fix the packet records as available runs from the 2025-02-18 KEV listing on an expedited tier, because the appliance being bypassed is the boundary the rest of the estate is built behind; a standard operating-system patch window measures this device by the wrong yardstick. Scope it to PAN-OS firewalls: the packet states the issue does not affect Cloud NGFW or Prisma Access software, so those deployments are not instances of this CVE and do not belong in the remediation population. Precondition, and this is where the control is routinely over-claimed: restricting the management interface bounds who can send the crafted request, it does not repair the path confusion. An attacker already positioned inside the permitted management segment, or on a jump host or admin workstation that legitimately reaches it, still invokes the restricted PHP scripts without authenticating. Isolation is the holding measure for the window before the fix lands, not a closure, and it is unavailable where the management interface must stay reachable for an out-of-band operational path. Distinguishing test: from a general user segment and from an external address, attempt to load the PAN-OS management web interface; anything that answers is within reach of the exploit the packet records as publicly available.",
+        "evidence": "Packet facts: name 'Palo Alto Networks PAN-OS Authentication Bypass Vulnerability', CWE-306; CISA KEV-listed 2025-02-18 with active_exploitation 'confirmed'; CVSS 9.1, RWEP 81; poc_available true; patch_available true, live_patch_available false. Vector states the flaw 'enables an unauthenticated attacker with network access to the management web interface to bypass the authentication otherwise required by the PAN-OS management web interface and invoke certain PHP scripts', that it 'can negatively impact integrity and confidentiality of PAN-OS', that risk is greatly reduced 'by restricting access to the management web interface to only trusted internal IP addresses', and that the issue 'does not affect Cloud NGFW or Prisma Access software'.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "NIST-800-53-SC-7"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-032",
+        "name": "PERIMETER-COMPROMISE-REBUILD-NOT-PATCH",
+        "description": "The packet is explicit that invoking the PHP scripts does not itself enable remote code execution, so the trigger for this control has to be stated precisely rather than assumed. Where a PAN-OS management web interface was reachable during the exposure window and the separate command-injection flaw the packet names — CVE-2024-9474 — was also unremediated, the device meets this control's condition exactly: the packet's attack_vector has the bypass commonly chained with that flaw to reach full remote code execution on a perimeter device under confirmed in-the-wild exploitation. For those units, applying the fix is not remediation. The response is configuration extraction for forensics, rebuild from a vendor image, and rotation of every secret the configuration held — local administrator credentials, RADIUS and TACACS+ shared secrets, IPsec pre-shared keys, API keys, and any certificate private material generated or stored on the device — because an attacker-written file, account, or script survives an upgrade untouched. Where only the bypass itself was reachable, the packet still records impact to the integrity and confidentiality of PAN-OS, which means the running configuration must be compared against a known-good copy and every secret that configuration stores treated as read, even if a full rebuild is not warranted. Precondition: this control is a response procedure, not a preventive one — it does nothing for a device that has not yet taken the fix, and rebuilding a device whose management interface is still broadly reachable simply restores an exploitable unit. Sequence it after the reachability restriction and the vendor fix, not instead of them.",
+        "evidence": "Packet facts: attack_vector states the attacker 'exploits an nginx/Apache path-confusion bug to invoke restricted PHP scripts without authenticating, then commonly chains the bypass with a separate command-injection flaw (CVE-2024-9474) to reach full remote code execution'; vector records impact to 'integrity and confidentiality of PAN-OS'; active_exploitation 'confirmed', CISA KEV 2025-02-18, poc_available true, RWEP 81, CVSS 9.1; patch_available true with live_patch_available false.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "UK-CAF-B4",
+          "ISO-27001-2022-A.8.9"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-031",
+        "name": "FIREWALL-EGRESS-TELEMETRY-SEPARATE-TRUST-ZONE",
+        "description": "What an exploit doing exactly what this packet describes emits is a request to the PAN-OS management web interface that reaches a restricted PHP script and returns successfully, with no preceding successful authentication event for that source or session — the authentication step is bypassed, not defeated. That is the signal to key on, and it lives in the management-plane web request log on the firewall itself. The control requires those logs, alongside the device's system and authentication logs, to be forwarded continuously to a collector in a separate trust zone: different management plane, different credentials, and an authentication path that does not depend on anything the firewall stores. Note what will not see this. There is no failed-login burst, no lockout, no crash and no reboot, so a detection built on authentication failures, availability loss, or administrator-session anomalies misses an attempt that behaves precisely as the packet describes. Preconditions, both load-bearing. The forwarding must already be configured and the collector already outside the device's blast radius before the attempt — a rule written after the fact against telemetry nobody was shipping produces nothing, and a syslog target whose credentials live in the firewall configuration is inside the same blast radius given the confidentiality impact the packet records. And the survival half of this control only bites where the chain to CVE-2024-9474 was reachable and the attacker held the device; for the bypass alone the value is visibility, not log survival. Detection prevents nothing here — it bounds the exposure window to alert-and-response time while the reachability restriction and the vendor fix are put in place.",
+        "evidence": "Packet facts: vector states an unauthenticated attacker with network access to the management web interface can 'bypass the authentication otherwise required by the PAN-OS management web interface and invoke certain PHP scripts' with impact to 'integrity and confidentiality of PAN-OS'; attack_vector records the chain with CVE-2024-9474 'to reach full remote code execution'; active_exploitation 'confirmed', CISA KEV 2025-02-18, poc_available true.",
+        "gap_closes": [
+          "NIS2-Art21-network-security"
+        ]
+      }
+    ]
   },
   "CVE-2024-53704": {
     "name": "SonicWall SonicOS SSLVPN Improper Authentication Vulnerability",
@@ -38996,7 +39421,39 @@
         "adequate": false,
         "gap": "The actual breach vector was a compromised vendor SaaS API key, not a direct customer-side flaw; CAF B4 (supply chain) requires modeling vendor-side compromise blast radius for privileged remote-access providers, which most customers had not done."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-032",
+        "name": "PERIMETER-COMPROMISE-REBUILD-NOT-PATCH",
+        "description": "BeyondTrust Privileged Remote Access and Remote Support exist to broker privileged sessions into other systems, and the packet's path ends with command execution on that broker itself — an unauthenticated crafted WebSocket-upgrade request whose Sec-WebSocket-Protocol / X-Ns-Company header values are concatenated unsanitized into a shell command run as the appliance's site user. For this product the control means an appliance whose HTTP surface was reachable during the exposure window is handled as attacker-held rather than as a patch item: configuration and session artifacts captured for forensics first, the appliance rebuilt from vendor media, and every credential the appliance stored, brokered, or that authenticated through it during the window rotated. Applying the vendor update closes the injection sink and removes nothing the attacker wrote or read through it, and on a privileged-access appliance what it read is the credential set for the downstream estate — which is why a remediation that ends at 'the fix is deployed' understates this CVE specifically. Precondition and limits: this is an incident-response default, not a preventive measure. It presumes the operator can bound an exposure window (the packet gives the 2024-12-19 KEV listing as the anchor and no per-unit indicator of compromise), and it does not tell an operator whether a given appliance was actually reached. Where session and access logs live only on the appliance the attacker had command execution on, absence of evidence is not evidence of absence, so reachability during the window — not a confirmed detection — has to be the trigger. Rebuild-and-rotate is also not free: it interrupts the remote-support function the appliance provides, which is the reason the patch-in-place shortcut gets taken.",
+        "evidence": "Packet: CWE-77 command injection; an unauthenticated attacker sends a crafted WebSocket-upgrade HTTP request with malicious Sec-WebSocket-Protocol / X-Ns-Company header values that are concatenated unsanitized into a shell command run as the appliance's site user, yielding remote command execution on the PAM appliance itself. CISA KEV-listed 2024-12-19; active_exploitation confirmed; poc_available true; CVSS 9.8; RWEP 74. patch_available true, live_patch_available false, with no live-patch note recorded — the vendor update is the only remediation the packet registers, and it is exactly the step this control says is insufficient on its own once exploitation is confirmed.",
+        "gap_closes": [
+          "NIST-800-53-SI-2",
+          "NIS2-Art21-vulnerability-handling"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-030",
+        "name": "PERIMETER-DEVICE-ZERODAY-SLA-TIER",
+        "description": "A Privileged Remote Access / Remote Support appliance is a remote-access trust boundary — it accepts connections from technicians outside the protected network — and the packet's request arrives unauthenticated over that same HTTP surface. So the remediation clock for this CVE cannot be the general application-patching SLA the AU Essential Eight and boundary-protection attestations are measured against: the requirement is the vendor fix deployed, or the vulnerable HTTP/WebSocket surface withdrawn from untrusted networks, measured from the 2024-12-19 KEV listing rather than from the next maintenance window. Precondition, and it is the whole difficulty on this product: withdrawing the surface is only available where remote technicians can still reach the appliance another way — an internal-only deployment, or an interposed VPN or authenticating proxy that terminates the connection before the appliance's own handler ever parses the upgrade headers. On an appliance published to the internet so external support staff can connect, which is the deployment this product is bought for, there is no isolation lever that preserves the function, and the vendor update is the only path; the isolation half then applies only to management and API surfaces that have no need to be publicly reachable. Note also what isolation does not do: it bounds who can send the crafted request, it does not repair the unsanitized concatenation, so any caller inside a permitted segment still reaches command execution. The packet records no live-patch mechanism, so there is no in-place option that avoids taking the vendor update.",
+        "evidence": "Packet: unauthenticated network request against the appliance's WebSocket-upgrade path yields remote command execution on the PAM appliance itself; CWE-77. CISA KEV-listed 2024-12-19; active_exploitation confirmed; poc_available true; CVSS 9.8; RWEP 74. patch_available true; live_patch_available false, with no live-patch note recorded.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "NIST-800-53-SC-7"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-046",
+        "name": "PEN-TEST-SCOPE-INCLUDES-SECURITY-PRODUCTS",
+        "description": "The vulnerable product here is itself a security control — the privileged remote access and remote support tier that brokers privileged sessions — which is precisely the class this control says must appear in test scope as attack surface rather than as a defense to work behind. Bound to this CVE that means the PRA/RS appliance's public HTTP surface, and specifically its WebSocket upgrade handling, is a tested surface in its own right: the defect lives in how attacker-supplied Sec-WebSocket-Protocol and X-Ns-Company header values are carried into a shell command, and no amount of testing the systems reached through the appliance would surface it. Scope language has to name the appliance and its API surface explicitly, because scope statements typically enumerate the estate the privileged-access tool protects and silently exclude the tool. Precondition and limits: this is an assurance control, not a mitigation. It changes the probability of finding the next defect in this surface and does nothing for an appliance already exposed to this one, and because the code is the vendor's, what it produces is a vendor report plus an operator exposure decision — not a fix the operator can write, which is the same reason a secure-coding attestation over the operator's own code passes cleanly while this path stays open. Keep the scope addition to the products the packet names, Privileged Remote Access and Remote Support; extending it to other security products is a decision to make on their own evidence, not on this entry.",
+        "evidence": "Packet: the affected product is BeyondTrust Privileged Remote Access (PRA) and Remote Support (RS) — a privileged-access product — and the flaw is a CWE-77 command injection reached by an unauthenticated crafted WebSocket-upgrade HTTP request through the Sec-WebSocket-Protocol / X-Ns-Company headers, executing as the appliance's site user. poc_available true; active_exploitation confirmed; CISA KEV-listed 2024-12-19; CVSS 9.8; RWEP 74.",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.28",
+          "UK-CAF-B4"
+        ]
+      }
+    ]
   },
   "CVE-2022-23227": {
     "name": "NUUO NVRmini2 Devices Missing Authentication Vulnerability",
@@ -39181,7 +39638,30 @@
         "adequate": false,
         "gap": "Application hardening (disabling unnecessary auto-execution features like Autorun-directory command execution) would have closed this vector even before the vendor patch existed."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-038",
+        "name": "MITIGATION-ACTIVE-PATCH-PENDING-VERDICT-CLASS",
+        "description": "The packet puts the flaw in a default: 'an unauthenticated user can import and execute arbitrary Bash or PowerShell commands on the host system by leveraging the default settings of the Autorun directory', on Cleo Harmony, VLTrader and LexiCom before 5.8.0.24. That gives a Cleo estate two states a compliance verdict must not collapse into one. The first is a host on 5.8.0.24 or later, where the defect is gone. The second is a host on an earlier build with the Autorun directory's auto-execution turned off by configuration — the execution trigger is closed, and the residual risk in that state is concrete for this product: the unauthenticated import path itself is untouched, so attacker-chosen files still land on the host, and the setting is a product default, so any operation that reapplies defaults (a configuration restore, a reinstall, an upgrade path that re-writes them) silently returns the host to full exposure with nothing in the version string to show it. The requirement is that the second state is reported as a distinct compensating-control state with a dated action item to reach the fixed build, and never as 'patched per SLA'. Distinguishing test: per Cleo host, report the product build and the Autorun auto-execution setting as two separate fields, and re-read the setting after every upgrade, restore and reinstall. An inventory whose only field is 'Cleo: remediated' cannot tell a 5.8.0.24 host from one whose mitigation was reverted by a default-restoring operation, and both read compliant. Precondition: disabling the auto-execution bounds the trigger — it does not remove the defect and it evicts nothing. The packet records active_exploitation as confirmed, so a host that ran a pre-5.8.0.24 build while reachable is not made clean by the configuration change; that host belongs on the response path, and the configuration state describes only what an attacker can do to it from this point forward.",
+        "evidence": "Packet: 'Cleo Multiple Products Unauthenticated File Upload Vulnerability', cwe_refs CWE-77, cvss 9.8, rwep_score 54, poc_available false. Vector: 'In Cleo Harmony before 5.8.0.24, VLTrader before 5.8.0.24, and LexiCom before 5.8.0.24, an unauthenticated user can import and execute arbitrary Bash or PowerShell commands on the host system by leveraging the default settings of the Autorun directory.' attack_vector: 'An unauthenticated attacker imports a file that lands in the Cleo product's default Autorun directory; the platform's default configuration automatically executes files placed there as Bash or PowerShell commands, giving the attacker code execution on the host with no credentials required.' cisa_kev true, kev_date 2024-12-17, active_exploitation confirmed. patch_available true; live_patch_available false (live_patch_notes null). Citing gaps closed here: NIST-800-53-CM-7 (Least Functionality), ISO-27001-2022-A.8.9 (Configuration management), AU-Essential-8-App-Hardening (User application hardening).",
+        "gap_closes": [
+          "NIST-800-53-CM-7",
+          "ISO-27001-2022-A.8.9",
+          "AU-Essential-8-App-Hardening"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-032",
+        "name": "PERIMETER-COMPROMISE-REBUILD-NOT-PATCH",
+        "description": "Cleo Harmony, VLTrader and LexiCom sit at the trust boundary by function — the packet's attack path requires that an unauthenticated party can reach an instance and import a file to it — and the packet's outcome is arbitrary Bash or PowerShell execution on the host with no credential required. With active exploitation recorded as confirmed, upgrading to 5.8.0.24 removes the import-to-execute path and removes nothing the executed commands already wrote or started. The control's default therefore binds here: for any instance that was reachable while running a build before 5.8.0.24, export the configuration for review, rebuild the host from a known-good image onto the fixed build rather than upgrading over the running state, and rotate every credential the instance held — a file-transfer platform's stored credentials reach outward to the partners it exchanges with, so the rotation scope does not stop at this host. This is what the flaw-remediation gap recorded on this entry is pointing at: the exposure window opened before a fix existed, so 'the host is on 5.8.0.24' answers what an attacker can do next and says nothing about what already happened. Precondition, stated rather than implied: rebuilding is a response action for an instance that was reachable during the exposure window. It prevents nothing, and it is not a substitute for the fixed build — a rebuilt instance that comes back on an earlier build is exposed again immediately, since the packet records the vendor fix as available and no live-patch path. It also reaches only artefacts on the instance itself; anything the executed commands pushed to another host is outside what rebuilding this one can undo and needs its own scope.",
+        "evidence": "Packet vector: 'In Cleo Harmony before 5.8.0.24, VLTrader before 5.8.0.24, and LexiCom before 5.8.0.24, an unauthenticated user can import and execute arbitrary Bash or PowerShell commands on the host system by leveraging the default settings of the Autorun directory.' attack_vector adds that the platform's default configuration 'automatically executes files placed there as Bash or PowerShell commands, giving the attacker code execution on the host with no credentials required'. cwe_refs CWE-77, cvss 9.8, rwep_score 54, poc_available false. cisa_kev true, kev_date 2024-12-17, active_exploitation confirmed. patch_available true; live_patch_available false (live_patch_notes null). Citing gaps closed here: NIS2-Art21-incident-handling (Incident handling) and NIST-800-53-SI-2 (Flaw Remediation).",
+        "gap_closes": [
+          "NIS2-Art21-incident-handling",
+          "NIST-800-53-SI-2"
+        ]
+      }
+    ]
   },
   "CVE-2024-35250": {
     "name": "Microsoft Windows Kernel-Mode Driver Untrusted Pointer Dereference Vulnerability",
@@ -41834,7 +42314,31 @@
         "adequate": false,
         "gap": "The KEV due date (2026-07-18) was three days after listing, against an exploitation window that opened six weeks after the May patch — a standard flaw-remediation SLA is far too slow for an unauthenticated ERP takeover already probed in the wild."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-129",
+        "name": "ENTERPRISE-ERP-MANAGEMENT-FUNCTION-AUTHENTICATION",
+        "description": "The packet puts this defect inside one endpoint of the ERP rather than in the surrounding account model: the Oracle Payments File Transmission endpoint (ibytransmit) performs no authentication, and an unauthenticated attacker with HTTP reach sends crafted requests that read arbitrary server files and end in takeover of the Payments component. Bound to this deployment the control means that endpoint — and every other EBS servlet that performs a privileged file or payment operation — makes its own authorization decision before it acts, instead of inheriting a verdict from whatever web tier or reverse proxy fronts it, and that no EBS instance in the affected 12.2.3 through 12.2.15 range is left with that endpoint reachable from a segment with no operational need for it, which for most estates means it should not answer from the internet at all. This is precisely why the identity-and-access-control gap recorded against this entry does not close the path: the attacker never holds an EBS account, so per-account privilege scoping is never consulted, and an attestation that every EBS user is uniquely identified and least-privileged passes cleanly while the endpoint stays open. Distinguishing test: from a segment with no operational need to reach Payments, issue unauthenticated HTTP requests to the File Transmission endpoint on a staging EBS instance and confirm each is refused before the file operation runs. Preconditions. The endpoint-side authorization is a property the vendor update establishes — this control states what to verify, it does not implement it, and restricting reachability in the meantime bounds who can send the request while leaving the endpoint fully exploitable to anything inside the permitted segment; that lever is unavailable entirely where the transmission endpoint must stay reachable for a payment counterparty. And because exploitation is confirmed and the primitive the packet names is an arbitrary file read, an instance that was reachable during the exposure window must have every credential, wallet, and key readable by the EBS application-tier account treated as disclosed and rotated: the update closes the path but returns nothing already read through it.",
+        "evidence": "Packet facts: name 'Oracle E-Business Suite Improper Privilege Management Vulnerability', CWE-269, CWE-287 and CWE-306. Vector: 'Vulnerability in the Oracle Payments product of Oracle E-Business Suite (component: File Transmission). Supported versions that are affected are 12.2.3-12.2.15. Easily exploitable vulnerability allows unauthenticated attacker with network access via HTTP to compromise Oracle Payments. Successful attacks of this vulnerability can result in takeover of Oracle Payments', CVSS 3.1 vector AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H. attack_vector: the attacker 'sends crafted HTTP requests to the Oracle Payments File Transmission (ibytransmit) endpoint, abusing missing authentication and improper privilege management to read arbitrary server files and ultimately take over the Payments component.' CISA KEV-listed 2026-07-15, active_exploitation 'confirmed'; patch_available true, live_patch_available false.",
+        "gap_closes": [
+          "UK-CAF-B2",
+          "NIST-800-53-SC-7"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-001",
+        "name": "CISA-KEV-RESPONSE-SLA",
+        "description": "The prioritisation trap on this entry is visible in its own numbers: CVSS 9.8 for an unauthenticated network takeover of a payments component, but RWEP 53, because the packet records no public proof-of-concept. A programme that ranks by exploit availability, or that treats an ERP patch as a quarterly suite activity, will schedule this behind items with public exploit code while the packet records exploitation as confirmed. The control's requirement here is that the 2026-07-15 KEV listing starts an expedited clock for every EBS instance in the affected 12.2.3 through 12.2.15 range, and that completion is measured per instance against the vendor's fixed level for the Payments component rather than by the presence of an approved patch in the change queue — an estate that carries 'Oracle EBS 12.2' as a single inventory row cannot tell which of its instances is still exposed. The remediation the packet records is the vendor update, with no live-patching primitive for this product and no reboot required, so the usual reason an ERP security fix slips — waiting for a restart window — is not available on this entry and should not be accepted as a deferral. Preconditions. With no live-patch path there are only two states in which the SLA is met: the fixed level is running, or documented compensating controls are in place with a dated action item — restricting who can reach the File Transmission endpoint is the compensating control, and it is a holding measure, not a closure. And the internet-facing instances belong at the front of the queue, because the packet's only stated access requirement is network access via HTTP.",
+        "evidence": "Packet facts: CISA KEV-listed 2026-07-15 with active_exploitation 'confirmed'; CVSS 9.8, RWEP 53; poc_available false; patch_available true, live_patch_available false, live_patch_notes 'No live-patching primitive for this product; the vendor update (no reboot required) is the remediation.' Vector records affected supported versions 12.2.3-12.2.15 and an 'unauthenticated attacker with network access via HTTP' able to compromise and take over Oracle Payments.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "ISO-27001-2022-A.8.8",
+          "NIST-800-53-SI-2",
+          "NIS2-Art21-vulnerability-management"
+        ]
+      }
+    ]
   },
   "CVE-2023-4346": {
     "name": "KNX Association KNX Protocol Connection Authorization Option 1 Overly Restrictive Account Lockout Mechanism Vulnerability",
@@ -42037,7 +42541,29 @@
         "adequate": false,
         "gap": "Perimeter boundary controls do not stop an SSRF that turns the trusted appliance itself into the request origin, letting outbound calls reach internal services the firewall would otherwise block."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-030",
+        "name": "PERIMETER-DEVICE-ZERODAY-SLA-TIER",
+        "description": "The SMA1000 Work Place interface is the appliance's published remote-access surface, and the packet places the defect there with a remote unauthenticated attacker able to coerce the appliance into issuing server-side requests to destinations of the attacker's choosing. Applied to this device, the control means the SMA1000 population runs on an expedited clock measured from the 2026-07-14 KEV listing to the moment each unit has both taken the vendor update and completed the reboot that update requires — not on the estate's standard appliance-patch window, because the device is the trust boundary and the flaw needs no credential to reach. The distinguishing test is per-unit: produce the running build and the last restart time and show the restart followed the update. An inventory that records the fixed release as 'applied' while the unit is still running pre-update code marks an exposed appliance compliant, and the packet registers no live-patch path that would let the fix take effect without that restart. Precondition, and it is why this control cannot fall back on the isolation alternative that normally bounds a perimeter-device defect: the vulnerable interface is the Work Place portal the appliance exists to publish to remote users, so 'restrict reachability of the affected interface' is unavailable except where the deployment can constrain the portal to known client source ranges. Where it cannot, there is no interim state — the update and its reboot are the only remediation, and the exposure runs until the restart completes. The absence of a public exploit is not a deferral argument here: the packet records no PoC available alongside confirmed in-the-wild exploitation, so the flaw is in use while defenders have nothing public to test their own exposure against.",
+        "evidence": "Packet entry 'SonicWall SMA1000 Appliances Server-Side Request Forgery Vulnerability' (CWE-918), cisa_kev true with kev_date 2026-07-14 and active_exploitation 'confirmed'; cvss 10, rwep_score 61, poc_available false, ai_discovered false. Vector: 'A Server-side request forgery (SSRF) vulnerability has been identified in the SMA1000 Appliance Work Place interface. A remote unauthenticated attacker could potentially cause the appliance to make requests to unintended location.' patch_available true, live_patch_available false, live_patch_notes: 'No live-patching primitive for this product; the vendor update requires a reboot and is the remediation.'",
+        "gap_closes": [
+          "NIST-800-53-SI-2",
+          "AU-ISM-1546",
+          "UK-CAF-B4"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-031",
+        "name": "FIREWALL-EGRESS-TELEMETRY-SEPARATE-TRUST-ZONE",
+        "description": "An SSRF leaves no artifact on the attacker's side of the connection — the only observable is the SMA1000 itself originating requests it has no operational reason to make. Bound to this appliance, the control means the unit's syslog and web-request logs are forwarded to a collector in a separate trust zone (different management plane, different credentials from the appliance), and that egress records are retained for every connection whose source is the appliance, so the coerced requests are visible from the network side rather than only from the box that was coerced. The detection keys on exactly the behaviour the packet documents: requests issued by the appliance to attacker-chosen destinations, which the packet enumerates as internal reconnaissance targets, cloud-metadata endpoints, and pivot destinations. It must not key on crashes, malware artifacts, or tool signatures — an SSRF that behaves precisely as the packet describes crashes nothing, writes nothing to the appliance's filesystem, and rides the appliance's own legitimate outbound HTTP client, so a rule built on those signals would miss every attempt. Precondition: the logs and egress records have to be flowing off-box before the attempt. A rule authored afterwards against telemetry nobody was collecting produces nothing, and appliance-local logs are the first thing lost when a unit is rebuilt during remediation. And this closes nothing on its own — it bounds exposure to alert-and-response time during the window before the vendor update and its reboot land, and it does not invalidate what was already retrieved: anything the appliance's own network position could reach, including the cloud-metadata material the packet names, stays valid after the patch until it is rotated.",
+        "evidence": "Packet attack_vector: 'A remote unauthenticated attacker sends a crafted request to the SMA1000 Work Place interface that coerces the appliance into making server-side requests to attacker-chosen destinations, enabling internal reconnaissance, cloud-metadata theft, or pivoting through the trusted appliance.' active_exploitation 'confirmed'; poc_available false. Citing gap ISO-27001-2022-A.8.16 (Monitoring activities) is recorded against this entry. live_patch_available false with live_patch_notes stating the vendor update requires a reboot and is the remediation.",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.16"
+        ]
+      }
+    ]
   },
   "CVE-2026-15410": {
     "name": "SonicWall SMA1000 Appliances Code Injection Vulnerability",
@@ -42291,7 +42817,30 @@
         "adequate": false,
         "gap": "The 2024-10-09 KEV due date trailed real exploitation; SI-2 flaw-remediation SLAs measured in weeks are far longer than the hours-to-days scanning window for an internet-facing preauth RCE."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-018",
+        "name": "SCANNER-PAPER-COMPLIANCE-TEST",
+        "description": "For HugeGraph-Server the fix is not a version number. The packet's own remediation is three simultaneous conditions — the 1.3.0 build, running on Java11, with the Auth system enabled — against an affected range that spans 1.0.0 before 1.3.0 on both Java8 and Java11. A scanner or CMDB that reports the installed HugeGraph version alone therefore marks as remediated a server sitting at the fixed build on a Java8 runtime, or at the fixed build with Auth left off, while the path the packet describes stays fully reachable: an unauthenticated crafted Gremlin script submitted to /apis/gremlin that abuses Java reflection/OGNL to escape HugeGraphSecurityManager into arbitrary OS commands. Applied to this product, the control means each HugeGraph instance produces three facts together — the running build, the JRE major version of the server process itself rather than the host's default java, and the Auth system's state read from live configuration rather than from a deployment template. The distinguishing test: stand up a staging instance at the fixed build with Auth disabled and submit an unauthenticated Gremlin script to /apis/gremlin; if it evaluates, the estate's version-only attestation is passing on an exploitable server. Precondition: this control tells you whether a server is in the fixed state, it does not put it there — the packet records the vendor update as the remediation, with no live-patching primitive for this product and no reboot required. And because active exploitation is confirmed and a public PoC exists, a server whose /apis/gremlin answered untrusted callers before remediation belongs on the triage path, not closed on the version check: reaching the fixed state removes the flaw, not what ran through it.",
+        "evidence": "Packet vector: 'RCE-Remote Command Execution vulnerability in Apache HugeGraph-Server. This issue affects Apache HugeGraph-Server: from 1.0.0 before 1.3.0 in Java8 & Java11. Users are recommended to upgrade to version 1.3.0 with Java11 & enable the Auth system, which fixes the issue.' Packet attack_vector: an unauthenticated attacker submits a crafted Gremlin script to /apis/gremlin that abuses Java reflection/OGNL to escape HugeGraphSecurityManager and run arbitrary OS commands on the server. CWE-284; CVSS 9.8; RWEP 65; poc_available true; CISA KEV-listed 2024-09-18 with active_exploitation confirmed. patch_available true, live_patch_available false, live_patch_notes: 'No live-patching primitive for this product; the vendor update (no reboot required) is the remediation.'",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "ISO-27001-2022-A.8.8",
+          "NIST-800-53-SI-2",
+          "NIS2-Art21-vulnerability-management"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-060",
+        "name": "DATABASE-SERVER-SIDE-SCRIPTING-DEFAULT-DENY",
+        "description": "The /apis/gremlin endpoint is a server-side scripting interface on a datastore: it accepts a caller-supplied script and evaluates it inside the server process, and the packet's exploit is that evaluation escaping HugeGraphSecurityManager through Java reflection/OGNL into arbitrary OS commands. This control's premise — that a datastore's server-side scripting capability is not simply live in the shipped posture, and that turning it on carries a documented decision — is what the packet's own remediation reaches for when it instructs operators to enable the Auth system: the deployments in scope are running a script-evaluation endpoint with no access decision in front of it. For HugeGraph the requirement cannot be stated as 'disable scripting', because Gremlin submission is how the product is queried; the deployable form is that no instance accepts script submission with the Auth system off, that network reachability of the endpoint is limited to callers with an operational need to query the graph, and that exposing an evaluation endpoint at all is recorded as an accepted risk with a named owner rather than inherited from the default configuration. Precondition, and this is the half the control is usually over-claimed on: the in-process sandbox is not the boundary — the packet's path is an escape from HugeGraphSecurityManager, so authentication and reachability bound who can submit a script, they do not make evaluation safe. Any caller who legitimately authenticates, and any host inside the permitted segment, still reaches the same primitive until the vendor update the packet records is applied.",
+        "evidence": "Packet attack_vector: an unauthenticated attacker submits a crafted Gremlin script to /apis/gremlin that abuses Java reflection/OGNL to escape HugeGraphSecurityManager and run arbitrary OS commands on the server. Packet vector records the recommended remediation as upgrading to 1.3.0 with Java11 and enabling the Auth system. CWE-284 (improper access control); CVSS 9.8; RWEP 65; poc_available true; CISA KEV-listed 2024-09-18, active_exploitation confirmed. patch_available true; live_patch_available false with the packet note that there is no live-patching primitive for this product and the vendor update (no reboot required) is the remediation.",
+        "gap_closes": [
+          "UK-CAF-B4"
+        ]
+      }
+    ]
   },
   "CVE-2014-0502": {
     "name": "Adobe Flash Player Double Free Vulnerability",
@@ -42880,7 +43429,39 @@
         "adequate": false,
         "gap": "Flaw remediation closed the access-control bug, but the observed ransomware window (<1h) and the credential-reuse follow-on mean patch-only remediation without a mandatory local-user password reset leaves the device compromisable."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-032",
+        "name": "PERIMETER-COMPROMISE-REBUILD-NOT-PATCH",
+        "description": "The firewall is the perimeter here, and the packet's chain does not end at the firewall bug: an unauthenticated remote attacker reaches SonicOS management access, and ransomware affiliates then use reused local-user credentials on SSL VPN to gain rapid internal footholds. That second half is what makes patch-in-place the wrong default for this device. Upgrading the firmware closes the improper-access-control path and leaves every local user account the appliance holds exactly as the attacker found it, and those credentials are the packet's own stated pivot into the estate. Bound to this product, the control means an appliance that was reachable on Gen 5 or Gen 6 firmware, or on Gen 7 running SonicOS 7.0.1-5035 or older, is handled as attacker-held rather than as a version-string problem: export the configuration for review, rebuild the unit onto the fixed firmware rather than upgrading over the running state, and rotate every local account the appliance stores, SSL VPN local users first, because that is the account class the packet names. Distinguishing test, and it is what separates this from paper closure: for each firewall produce the installed firmware version alongside the date each local user's password was last set, and confirm no account still carries a password that predates the upgrade. An appliance that reports the fixed build while its SSL VPN local accounts keep their pre-upgrade passwords is still exploitable by the exact chain the packet describes, and both the patch-application and flaw-remediation attestations read clean on it. Preconditions: this is the response path for a unit that was reachable during the exposure window — it prevents nothing on its own, and it does not replace the vendor update, which the packet records as the remediation with no live-patch path and a required reboot, so a rebuilt unit that has not come back on the fixed firmware is exposed again the moment it is online. Rotation also reaches only credentials this appliance holds; where the same passwords were reused on other systems, those systems are outside what rebuilding this device can fix and need their own rotation scope.",
+        "evidence": "Packet: 'SonicWall SonicOS Improper Access Control Vulnerability', cwe_refs CWE-284, cvss 9.8, rwep_score 62, poc_available false. Vector: 'An improper access control vulnerability has been identified in the SonicWall SonicOS management access, potentially leading to unauthorized resource access and in specific conditions, causing the firewall to crash. This issue affects SonicWall Firewall Gen 5 and Gen 6 devices, as well as Gen 7 devices running SonicOS 7.0.1-5035 and older versions.' attack_vector: 'An unauthenticated remote attacker reaches SonicOS management access to obtain unauthorized resource access (or crash the firewall); ransomware affiliates chained this with reused local-user credentials on SSL VPN to gain rapid internal footholds.' cisa_kev true, kev_date 2024-09-09, active_exploitation confirmed. patch_available true, live_patch_available false, live_patch_notes: 'No live-patching primitive for this product; the vendor update requires a reboot and is the remediation.' Citing gaps closed here: NIST-800-53-SI-2 (Flaw Remediation), AU-Essential-8-Patch (Patch operating systems), UK-CAF-D1 (Response and recovery planning).",
+        "gap_closes": [
+          "NIST-800-53-SI-2",
+          "AU-Essential-8-Patch",
+          "UK-CAF-D1"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-030",
+        "name": "PERIMETER-DEVICE-ZERODAY-SLA-TIER",
+        "description": "A SonicWall firewall is the device class this tier exists for, and the packet places the defect on its management access — pre-authentication, reachable by a remote attacker holding no credential, on the box that is itself the trust boundary. For this entry the tier means the clock runs from the 2024-09-09 KEV listing rather than from the next appliance maintenance window, and that where the fixed firmware cannot be taken inside that window the alternative is not a risk acceptance but isolation of the vulnerable interface: SonicOS management answers only from a named operator source or an out-of-band path, never from the internet. The reason the ordinary schedule does not fit this device is in the packet's own vector — the flaw can crash the firewall as well as expose management resources, so the failure mode of deferring includes losing the perimeter itself, and the attack_vector records affiliates converting the access into rapid internal footholds. Precondition, and it must be stated because this is exactly where the control gets over-claimed on firewalls: restricting reachability bounds who can send the request, it does not remove the defect, and it does not cover the whole chain the packet describes. Management access can usually be pulled back to an operator segment. The SSL VPN the packet names as the pivot cannot be — it exists to answer remote clients, and the credential-reuse step needs no access to the management interface at all, so a unit whose management plane has been restricted is still reachable through the half of the chain that turns valid local credentials into a foothold. The packet records the vendor update as the remediation, with no live-patch path and a reboot required, so interface restriction is the bridge to a reboot that still has to be scheduled, never a substitute for it.",
+        "evidence": "Packet: cisa_kev true with kev_date 2024-09-09, active_exploitation confirmed, cvss 9.8, rwep_score 62, cwe_refs CWE-284. Vector states the flaw is in 'the SonicWall SonicOS management access, potentially leading to unauthorized resource access and in specific conditions, causing the firewall to crash', affecting 'SonicWall Firewall Gen 5 and Gen 6 devices, as well as Gen 7 devices running SonicOS 7.0.1-5035 and older versions'. attack_vector: 'An unauthenticated remote attacker reaches SonicOS management access ... ransomware affiliates chained this with reused local-user credentials on SSL VPN to gain rapid internal footholds.' patch_available true, live_patch_available false, live_patch_notes: 'No live-patching primitive for this product; the vendor update requires a reboot and is the remediation.' Citing gaps closed here: NIS2-Art21-patch-management (Vulnerability handling and disclosure) and NIST-800-53-SC-7 (Boundary Protection).",
+        "gap_closes": [
+          "NIS2-Art21-patch-management",
+          "NIST-800-53-SC-7"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-031",
+        "name": "FIREWALL-EGRESS-TELEMETRY-SEPARATE-TRUST-ZONE",
+        "description": "This control names firewalls specifically, and the packet gives two reasons it binds to this one. First, the evidence that matters on this chain is authentication data the firewall itself produces: the packet's attack_vector is reused local-user credentials on SSL VPN, so the signal is a SUCCESSFUL SSL VPN authentication on a local firewall account from a source that account has not used before, followed by internal activity — not a failed-login pattern and not a malware artefact, because on this chain the credential is valid and the login is accepted. Second, the packet's vector states the flaw can cause the firewall to crash, and a crashed or restarted appliance is precisely when on-box log retention is thinnest. Applied here, the requirement is that SonicOS syslog, authentication and VPN session logs land on a collector in a separate trust zone — different management plane, different credentials, not reachable by administering the firewall — so the record of that login survives both an attacker holding the device and the device restarting. The distinguishing test is the alert, not the pipeline: replay a successful SSL VPN authentication for a local account from an unfamiliar source and confirm the collector raises an alert. A SIEM that ingests the firewall's logs but alerts only on authentication failures and denied traffic sees nothing at all on this chain, because every step the packet describes uses credentials the device accepts. Precondition: this preserves and surfaces evidence and prevents nothing — it does not bound reachability of SonicOS management access and it does not remediate the flaw, which the packet records as requiring the vendor update and a reboot. Its value is confined to the window between an attacker reaching the device and the estate responding, and it is worth nothing if the collector authenticates through the same appliance or the same credential set the attacker has just obtained.",
+        "evidence": "Packet vector: '... potentially leading to unauthorized resource access and in specific conditions, causing the firewall to crash.' attack_vector: 'An unauthenticated remote attacker reaches SonicOS management access to obtain unauthorized resource access (or crash the firewall); ransomware affiliates chained this with reused local-user credentials on SSL VPN to gain rapid internal footholds.' cisa_kev true, kev_date 2024-09-09, active_exploitation confirmed, poc_available false, cvss 9.8, rwep_score 62, cwe_refs CWE-284. patch_available true, live_patch_available false, live_patch_notes: 'No live-patching primitive for this product; the vendor update requires a reboot and is the remediation.' UK-CAF-D1 (Response and recovery planning) is recorded as an insufficient framework control on this entry.",
+        "gap_closes": [
+          "UK-CAF-D1"
+        ]
+      }
+    ]
   },
   "CVE-2017-1000253": {
     "name": "Linux Kernel PIE Stack Buffer Corruption Vulnerability",
@@ -42988,7 +43569,29 @@
         "adequate": false,
         "gap": "Malicious-code protection that signature-scans uploads does not catch a crafted MVG/MSL image whose payload is a delegate command-injection string rather than known malware, so the file reaches the vulnerable ImageMagick coder."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-144",
+        "name": "EMBEDDED-MEDIA-PARSER-LIBRARY-INVENTORY-AND-PATCH-PARITY",
+        "description": "ImageMagick is reached here as a component, not as an application anyone administers: the packet's path is a server handing an uploaded file to ImageMagick's delegate command, so the exposed copies are wherever that binary or library sits behind an image-accepting feature. The packet also names two separate fixed lines — 6.9.3-10 on the 6.x branch and 7.0.1-1 on 7.x — which is the parity trap for this CVE: a report reading 'ImageMagick updated' does not say which branch each copy is on, and a 6.x copy is not remediated by the existence of a 7.x fix. The control here means enumerating every ImageMagick copy the estate actually executes — distribution packages, copies baked into container images, and builds shipped inside an application bundle that no host package manager owns — and showing each at or above the fixed release for its own branch, rather than closing the flaw-remediation ticket when the one inventoried package updates. Scope it to ImageMagick: the packet ties these eight coders to ImageMagick and provides no mapping into other image-processing software, so instructing operators to treat every image-handling binary in the estate as an instance of this CVE manufactures findings and real removal work against software no evidence implicates. Widen only where a verified source identifies another product carrying the same coders. Precondition: this reaches only copies the inventory can see and the operator can rebuild — a copy inside a container image is remediated by rebuilding the image, not by updating the host package, and it returns silently on the next deploy of an unrebuilt image. And the two fixed releases the packet names are the floor for this defect, not a destination: an estate that pins a copy at that build and calls it compliant is holding a component fixed against this CVE alone.",
+        "evidence": "Packet vector: 'The (1) EPHEMERAL, (2) HTTPS, (3) MVG, (4) MSL, (5) TEXT, (6) SHOW, (7) WIN, and (8) PLT coders in ImageMagick before 6.9.3-10 and 7.x before 7.0.1-1 allow remote attackers to execute arbitrary code via shell metacharacters in a crafted image, aka \"ImageTragick.\"' Packet attack_vector: an attacker uploads a crafted MVG/MSL/SVG image whose coder fields contain shell metacharacters; when a server passes the file to ImageMagick's delegate command, the metacharacters are executed as an OS command, yielding remote code execution. CWE-20; CVSS 8.4; RWEP 72; poc_available true; CISA KEV-listed 2024-09-09 with active_exploitation confirmed. patch_available true; live_patch_available false, live_patch_notes: 'No live-patching primitive for this product; the vendor update (no reboot required) is the remediation.'",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.8",
+          "UK-CAF-B4"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-025",
+        "name": "CONFIGURATION-SIDE-LIVE-PATCH-INVENTORY",
+        "description": "This CVE has a configuration-side mitigation path that does not wait on any package update, and the packet names its surface precisely: the eight coders — EPHEMERAL, HTTPS, MVG, MSL, TEXT, SHOW, WIN, PLT — that reach a delegate command. ImageMagick's security policy configuration (policy.xml) can deny those coders on a copy that is still on a vulnerable build, so the control's requirement is that this path be inventoried, tested and deployable for every ImageMagick copy independently of when each copy's update lands, rather than researched during the incident. It matters more than usual here because of how delivery works: the exploit file is itself a legitimate image format the packet names, so an upload filter that accepts images accepts the payload, and the payload is text carrying shell metacharacters rather than a binary that a malicious-code engine has anything to match on. The distinguishing test keys on the behaviour the packet documents: submit a crafted MVG file whose coder fields carry shell metacharacters through the real upload path on a staging instance and confirm no delegate command runs — confirming that a policy file exists on disk proves nothing about the process that did the decoding, which may be a different ImageMagick copy or one pointed at another configuration directory by its environment. Preconditions, none of which this control removes: denying these coders is a functional change that breaks any workflow legitimately rendering them, so it needs a rendering test before rollout; it must be applied to every copy the upload path can reach, since one host's policy does nothing for a copy inside a container image; and it stops new attempts only — the packet records confirmed active exploitation with a public PoC, so a server that accepted uploads while exposed needs triage, because the policy evicts nothing already executed. The packet's vendor update remains the remediation.",
+        "evidence": "Packet vector names the eight vulnerable coders — EPHEMERAL, HTTPS, MVG, MSL, TEXT, SHOW, WIN, PLT — in ImageMagick before 6.9.3-10 and 7.x before 7.0.1-1, executing arbitrary code via shell metacharacters in a crafted image (aka 'ImageTragick'). Packet attack_vector: an uploaded crafted MVG/MSL/SVG image whose coder fields contain shell metacharacters is executed as an OS command when the server passes the file to ImageMagick's delegate command. CWE-20; CVSS 8.4; RWEP 72; poc_available true; CISA KEV-listed 2024-09-09, active_exploitation confirmed. patch_available true; live_patch_available false with the packet note of no live-patching primitive for this product and the vendor update (no reboot required) as the remediation.",
+        "gap_closes": [
+          "AU-Essential-8-App-Hardening",
+          "NIST-800-53-SI-3"
+        ]
+      }
+    ]
   },
   "CVE-2024-7262": {
     "name": "Kingsoft WPS Office Path Traversal Vulnerability",
@@ -43843,7 +44446,31 @@
         "adequate": false,
         "gap": "Configuring Microsoft Office to block internet-origin macros is the compensating control that neutralizes this file-borne RCE, but it is only effective where the baseline is actually enforced."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-120",
+        "name": "DOWNLOAD-PROVENANCE-ENFORCEMENT",
+        "description": "The packet conditions this exploit on delivery: an attacker-supplied Microsoft Project file, opened by a victim 'with internet-macro blocking disabled', yielding code execution in the user's context. That condition is what makes provenance the enforceable control here, and it is also what makes 'the policy is enabled' an inadequate attestation — the block-macros-from-the-internet safeguard is a provenance decision, so it only fires on a file the host recognises as externally sourced. Bound to this product, the requirement is that a Project file arriving by mail, web download or an untrusted file share still carries the Mark-of-the-Web when Project opens it, and that the mark survives the container it arrived in: a .mpp extracted from an archive by a tool that does not propagate the mark, mounted from an ISO or VHD, renamed, or re-saved onto a share the estate treats as trusted reaches Project as local content, and the safeguard the packet names as the missing one is never consulted. Scope it to what the packet establishes: it names Microsoft Project, so the population to enumerate is the Project installs and the macro and Trust Center policy state applied to them — this is not a basis for re-baselining every application in the estate. The distinguishing test: mail a Project file that would trigger the internet-macro block inside a ZIP to a managed workstation through the normal ingress path, extract it the way a user would, open it, and confirm the block still applies to the extracted copy. An application-hardening attestation that records the policy as Enabled in Group Policy passes cleanly while a provenance-stripped .mpp reaches the same parser with macro execution permitted. Precondition: this is a delivery-path control. It bounds which files reach the CWE-20 defect with macros allowed; it does not repair the defect, and it gives nothing against a file delivered through a path the estate has already marked trusted. The packet records a vendor patch as the remediation, with no live-patch path and no reboot required, which is the argument for treating provenance enforcement as the bridge to that update rather than as the fix.",
+        "evidence": "Packet: 'Microsoft Project Remote Code Execution Vulnerability', cwe_refs CWE-20, cvss 8.8, rwep_score 48, poc_available false. Vector: 'Microsoft Project Remote Code Execution Vulnerability'. attack_vector: 'An attacker delivers a malicious Microsoft Project file; when the victim opens it (with internet-macro blocking disabled) the flaw yields code execution in the user's context.' cisa_kev true, kev_date 2024-08-13, active_exploitation confirmed. patch_available true, live_patch_available false, live_patch_notes: 'No live-patching primitive for this product; the vendor update (no reboot required) is the remediation.' Citing gaps closed here: AU-Essential-8-App-Hardening (User application hardening), NIST-800-53-CM-7 (Least Functionality), UK-CAF-B4 (System security).",
+        "gap_closes": [
+          "AU-Essential-8-App-Hardening",
+          "NIST-800-53-CM-7",
+          "UK-CAF-B4"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-001",
+        "name": "CISA-KEV-RESPONSE-SLA",
+        "description": "The packet supplies the two facts that set this clock: a KEV listing on 2024-08-13 with confirmed exploitation, and a vendor update the entry records as the remediation with no reboot required. The no-reboot property is the operative one here — the usual reason an Office-component update slides into the next monthly maintenance window does not apply, so the clock runs from the KEV listing and completion is measured per install, against the build Microsoft Project actually reports, rather than against 'approved' or 'downloaded' in the management console. The specific way this remediation goes wrong on this entry is the compensating control being mistaken for the remediation. The packet's exploitation condition is that internet-macro blocking was disabled; that describes the estate where the attack succeeded and is not a statement that an estate with the policy enabled is unexploitable, because the CWE-20 defect stays in the parser for any Project file that reaches a user without an internet-origin mark. Enabling the policy buys the time to deploy; recording it as closure of a KEV item leaves the flaw in place under a clean attestation. Precondition: this control governs only installs the update path can actually reach, and it presumes the fixed build is deployable to them. Where a Project install cannot take the update, the state is the macro policy as an active mitigation and it has to be carried as an open, dated item, not as an SLA that was met.",
+        "evidence": "Packet: 'Microsoft Project Remote Code Execution Vulnerability', cwe_refs CWE-20, cisa_kev true, kev_date 2024-08-13, active_exploitation confirmed, cvss 8.8, rwep_score 48, poc_available false. attack_vector: 'An attacker delivers a malicious Microsoft Project file; when the victim opens it (with internet-macro blocking disabled) the flaw yields code execution in the user's context.' patch_available true, live_patch_available false, live_patch_notes: 'No live-patching primitive for this product; the vendor update (no reboot required) is the remediation.' Citing gaps closed here: NIST-800-53-SI-2 (Flaw Remediation), ISO-27001-2022-A.8.8 (Management of technical vulnerabilities), NIS2-Art21-vulnerability-management (Vulnerability handling).",
+        "gap_closes": [
+          "NIST-800-53-SI-2",
+          "ISO-27001-2022-A.8.8",
+          "NIS2-Art21-vulnerability-management"
+        ]
+      }
+    ]
   },
   "CVE-2024-32113": {
     "name": "Apache OFBiz Path Traversal Vulnerability",
@@ -44168,7 +44795,33 @@
         "adequate": false,
         "gap": "Internet-facing Now Platform instances present the vulnerable UI macros pre-authentication, so boundary controls that assume login-gated access do not apply."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-001",
+        "name": "CISA-KEV-RESPONSE-SLA",
+        "description": "The packet splits this estate in a way a single patch SLA does not model: ServiceNow applied the update to hosted instances and released it to partners and self-hosted customers, and names specific patches and hot fixes as what addresses the vulnerability. So the operator's obligation on the KEV clock that opened 2024-07-29 is three different actions depending on who runs the instance. On a vendor-hosted instance the operator cannot apply anything and the obligation is to obtain assurance that the instance actually carries the fix — an item to raise with the supplier and record, not an assumption to inherit from a status page. On a self-hosted instance it is applying the named patch or hot fix. On a partner-run instance it is the operator's own item to chase the partner to completion, because the exposure sits on the operator's data whoever holds the console. What makes the compressed clock enforceable here rather than aspirational is a packet fact: there is no live-patching primitive for this product and the vendor update requires no reboot, so nothing about downtime stands between the operator and remediation — the real delay is instance inventory and partner coordination, and both are addressable before the next disclosure rather than during it. Priority follows the packet: an unauthenticated path to code execution in the context of the Now Platform, a public PoC, confirmed in-the-wild exploitation and RWEP 74. The completion measure is per instance — every Now Platform instance the organization has, including any a partner operates on its behalf — showing the named patch or hot fix applied, not a count of instances 'on a current release'.",
+        "evidence": "Packet: CISA KEV-listed 2024-07-29, active_exploitation 'confirmed', poc_available true, CVSS 9.3, RWEP 74. Vendor description in the packet: the input-validation vulnerability was identified in Vancouver and Washington DC Now Platform releases; it 'could enable an unauthenticated user to remotely execute code within the context of the Now Platform'; 'ServiceNow applied an update to hosted instances, and ServiceNow released the update to our partners and self-hosted customers'; and the packet refers to the patches and hot fixes that address the vulnerability. patch_available true; live_patch_available false, with the packet's note stating there is no live-patching primitive for this product and that the vendor update (no reboot required) is the remediation.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "ISO-27001-2022-A.8.8",
+          "NIST-800-53-SI-2",
+          "NIS2-Art21-vulnerability-management",
+          "DORA-Art-9"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-018",
+        "name": "SCANNER-PAPER-COMPLIANCE-TEST",
+        "description": "Paper compliance on this entry has a precise shape, and it comes straight from the packet's wording: the vulnerability was identified in the Vancouver and Washington DC releases, and specific patches and hot fixes address it. A record showing an instance is on one of those named families therefore says nothing about whether the fix is present — the family release is not the unit of remediation here, the named patch or hot fix is, and a vulnerability-management report that closes this item on 'the instance is current' has recorded the vendor's release train rather than the presence of the fix. External scanning does not rescue it either, because the operator is looking at a hosted web application with no build banner to read. The operational test is behavioural and matches the path the packet documents: on a non-production instance, submit a Jelly/Glide template expression to a public, unauthenticated UI page and confirm it comes back as inert text rather than being evaluated server-side. That is exactly what an exploit doing what the packet describes would do — an unauthenticated submission to a public UI macro that the platform evaluates — so a probe that is refused or returned uninterpreted is evidence about the actual defect, whereas an instance-version report is evidence about a label. Preconditions: the probe must run against a non-production instance or under terms the vendor has agreed, since firing crafted input at a live production instance is an exploitation attempt in its own right; and on a vendor-hosted instance the operator may have no instance they are permitted to test, in which case this control degrades to a supplier-assurance question — which is the honest answer, not a reason to fall back on the release-train report.",
+        "evidence": "Packet attack vector: 'An unauthenticated attacker submits a Jelly/Glide template expression to a public UI macro; the Now Platform evaluates it server-side, giving arbitrary expression/code execution and access to instance data.' Packet vendor text names the Vancouver and Washington DC Now Platform releases and refers to the patches and hot fixes that address the vulnerability. CISA KEV-listed 2024-07-29, active_exploitation 'confirmed', poc_available true, CWE-1287.",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.8",
+          "NIST-800-53-SI-2",
+          "UK-CAF-B4"
+        ]
+      }
+    ]
   },
   "CVE-2024-39891": {
     "name": "Twilio Authy Information Disclosure Vulnerability",
@@ -45516,7 +46169,21 @@
         "adequate": false,
         "gap": "Technical-vulnerability management often omits embedded Chromium runtimes (Electron apps, WebView) that inherit the same V8 flaw but do not auto-update with the browser."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-057",
+        "name": "BROWSER-MANAGED-UPDATE-NO-DEFERRAL",
+        "description": "The packet's exposure is ordinary browsing: a victim visits an attacker-controlled or compromised page whose crafted JavaScript triggers an out-of-bounds write in V8, corrupting renderer memory, which chained with a sandbox escape yields code execution on the host. There is no user decision to harden and no attachment to quarantine — every managed endpoint running Google Chrome is in the exposed population the moment such a page loads, which is why the update clock is the lever with real leverage on this entry. For this CVE the control means the estate's Chrome update ring carries the fixed build inside the no-deferral window rather than a weekly or monthly ring, with completion measured by the version each browser reports rather than by the update policy being set in a console. The packet is unusually favourable on cost here and that removes the usual excuse: it records no live-patching primitive for this product and states the vendor update is the remediation with no reboot required, so there is no maintenance window to negotiate and no user-visible outage to schedule — a deferral on this entry buys nothing and is pure exposure. What still needs verifying is the browser that never closes: the update lands without a reboot, but a Chrome process that has been running since before it keeps executing the V8 it already loaded, so the version the running browser reports after relaunch is the measurement that counts, and long-uptime machines are the population to check first. Distinguishing test: enumerate the version actually reported by the running browser across a sample of the estate against the elapsed time since the vendor build shipped, weighted toward long-uptime hosts; an update-ring policy screenshot and a management-console compliant count are attestations about configuration, not evidence that the vulnerable V8 stopped executing anywhere. Scope this to Google Chrome — the packet names Chrome builds prior to 124.0.6367.207 and gives no mapping from this V8 into any other product, so treating every browser or embedded runtime in the estate as an instance of this CVE manufactures removal and update work against software this entry does not implicate. Precondition: a no-deferral ring compresses the window between the vendor build existing and the fleet running it. It does nothing for the window before the build existed, and the packet records confirmed in-the-wild exploitation with the KEV listing on 2024-05-16 — endpoints browsing before the update reached them were exposed to a bug already in use, so a host showing other indications of compromise belongs on the incident path rather than being closed on a version number.",
+        "evidence": "Packet: out-of-bounds write (CWE-787) in V8 in Google Chrome prior to 124.0.6367.207, reachable from a crafted HTML page; attack_vector describes a victim visiting an attacker-controlled or compromised page whose crafted JavaScript corrupts renderer memory, chained with a sandbox escape for code execution on the host. CISA KEV-listed 2024-05-16, active_exploitation confirmed, poc_available false, CVSS 8.8, RWEP 54. patch_available true; live_patch_available false, live_patch_notes: no live-patching primitive for this product, the vendor update (no reboot required) is the remediation. Citing gaps include AU-Essential-8-Patch, ISO-27001-2022-A.8.8, NIST-800-53-SI-2 and NIS2-Art21-vulnerability-management.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "ISO-27001-2022-A.8.8",
+          "NIST-800-53-SI-2",
+          "NIS2-Art21-vulnerability-management"
+        ]
+      }
+    ]
   },
   "CVE-2021-40655": {
     "name": "D-Link DIR-605 Router Information Disclosure Vulnerability",
@@ -47571,7 +48238,30 @@
         "adequate": false,
         "gap": "System-security assurance for a remote-access gateway is undermined when a mitigation ships and is then bypassed; point-in-time hardening does not survive an evolving exploit chain."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-030",
+        "name": "PERIMETER-DEVICE-ZERODAY-SLA-TIER",
+        "description": "Ivanti Connect Secure (9.x, 22.x), Ivanti Policy Secure (9.x, 22.x) and Ivanti Neurons for ZTA are the remote-access trust boundary, and the packet has an unauthenticated attacker reaching the SAML component, coercing the appliance into a server-side request to its own loopback backend on 127.0.0.1:8090, and chaining CVE-2024-21887 from there to run OS commands and drop webshells. For these products the control means the remediation clock starts at the 2024-01-31 KEV listing and is measured in hours, not folded into the next appliance maintenance window, because the device being remediated is the thing that decides who gets inside. The packet is specific about what completion means: there is no vendor live-patch mechanism and remediation requires applying the fixed release and rebooting, so a unit that has taken the fixed release but not rebooted still runs the vulnerable code and must be counted as exposed — and on a VPN concentrator that reboot is the step most likely to slip, because taking it drops every active remote-access session, which is exactly how a deferral gets recorded as 'patched'. Precondition, and this is where the control is usually over-claimed: its alternative to the compressed clock is isolating the vulnerable interface, and here that lever is largely unavailable, because the SSRF is reached through the SAML component of the appliance's public-facing web interface — the service itself. Restricting reachability of that interface means denying remote access to the workforce, so it is a time-boxed business decision to be taken explicitly, not an assumed compensating control. Nor does network boundary filtering substitute: the coerced request originates on the appliance and terminates on its own loopback address, so no ACL between the appliance and any other host is on that path.",
+        "evidence": "Packet vector: SSRF (CWE-918) in the SAML component of Ivanti Connect Secure (9.x, 22.x), Ivanti Policy Secure (9.x, 22.x) and Ivanti Neurons for ZTA, allowing access to restricted resources without authentication. Attack vector: forces a server-side request to the appliance's own loopback backend on 127.0.0.1:8090, then chains CVE-2024-21887 to run OS commands and drop webshells. CISA KEV listed 2024-01-31; active_exploitation confirmed; poc_available true; RWEP 84; CVSS 8.2; patch_available true; live_patch_available false; live_patch_notes: 'No vendor live-patch mechanism; remediation requires applying the fixed release and rebooting.' Citing gaps AU-Essential-8-Patch, NIST-800-53-SI-2.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "NIST-800-53-SI-2"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-032",
+        "name": "PERIMETER-COMPROMISE-REBUILD-NOT-PATCH",
+        "description": "The packet's chain does not stop at information disclosure: it ends in OS command execution and webshells on the appliance, with confirmed in-the-wild exploitation and a public PoC running from the 2024-01-31 KEV listing. For any Connect Secure, Policy Secure or Neurons for ZTA unit that was reachable during that window, applying the fixed release is therefore not remediation — it closes the SSRF into 127.0.0.1:8090 and the chained command-injection path, and removes nothing the attacker already placed on the box, including a webshell that keeps working across the upgrade because it is attacker-written content rather than vendor code. Bound to these appliances, the control means the default incident path is configuration and evidence capture, rebuild from vendor media, and rotation of every credential the appliance stored or brokered, in place of patch-in-place. The technical-vulnerability and system-security attestations cited on this entry close on the version the appliance reports, which is exactly the state a compromised-then-patched unit is in — they read clean while an implant persists — and the ICT risk-management obligation cited here inherits the same defect when its protection measures rest on a device that may already be attacker-held. Distinguishing test: for each unit, produce the date the fixed release was applied, the date it was rebuilt, and the credential-rotation record; a fleet where the first date exists and the second does not has patched, not recovered. Precondition: rebuild returns the unit to a known-good build and nothing more — it does not establish what was read or taken during the exposure window, so the rotation half is what bounds continued access, and any account that authenticated through the appliance while it was exposed has to be treated as exposed with it.",
+        "evidence": "Packet attack vector: the unauthenticated SSRF chains CVE-2024-21887 'to run OS commands and drop webshells' on the appliance. active_exploitation confirmed; poc_available true; CISA KEV listed 2024-01-31; RWEP 84; patch_available true; live_patch_available false; live_patch_notes: 'No vendor live-patch mechanism; remediation requires applying the fixed release and rebooting.' Citing gaps ISO-27001-2022-A.8.8, UK-CAF-B4, DORA-Art-9.",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.8",
+          "UK-CAF-B4",
+          "DORA-Art-9"
+        ]
+      }
+    ]
   },
   "CVE-2023-22527": {
     "name": "Atlassian Confluence Data Center and Server Template Injection Vulnerability",
@@ -50963,7 +51653,39 @@
         "adequate": false,
         "gap": "Technical-vulnerability-management closes on the CVE ticket without verifying the management-plane exposure that makes exploitation trivial was actually removed."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-030",
+        "name": "PERIMETER-DEVICE-ZERODAY-SLA-TIER",
+        "description": "BIG-IP is the edge load-balancer class this control names, and the packet locates the reachable surface precisely: the Configuration utility, reached through the BIG-IP management port and/or self IP addresses. For this CVE the tier means the remediation clock runs from the 2023-10-31 KEV listing, and the alternative to the fix inside that clock is withdrawing the Configuration utility from every network that is not an administrator network — self IPs in particular, because a self IP can be reachable from data-plane segments the management port is not, and an estate that has hardened only the management port has not withdrawn the surface the packet names. The packet records no vendor live-patch mechanism and states that remediation requires applying the fixed release and rebooting, so a unit with the release staged but not rebooted still carries the vulnerable code and must not be counted as remediated; on a load-balancer carrying production traffic that reboot is the step most likely to be deferred, and a deferral recorded as patched is the specific way this remediation goes wrong. Precondition, which is where this control is usually over-claimed: restricting reach to the Configuration utility bounds who can present the injected SQL, it does not repair the injection — and the packet pairs this CVE with CVE-2023-46747 to supply the unauthenticated reach, so an attacker who arrives through that bypass, or from a permitted administrator segment, still reaches arbitrary system-command execution. One further packet fact constrains the fix itself: the vector states that software versions which have reached End of Technical Support are not evaluated, so a unit on such a version has no evaluated fixed release to reach and its remediation is moving onto a supported branch rather than applying a hotfix to the branch it is on.",
+        "evidence": "Packet: CWE-89; an authenticated attacker with network access to the Configuration utility through the BIG-IP management port and/or self IP addresses executes arbitrary system commands, in practice paired with CVE-2023-46747 to gain the needed unauthenticated reach. CISA KEV-listed 2023-10-31; active_exploitation confirmed; poc_available true; CVSS 8.8; RWEP 75. patch_available true; live_patch_available false with the note 'No vendor live-patch mechanism; remediation requires applying the fixed release and rebooting.' The vector also states: 'Software versions which have reached End of Technical Support (EoTS) are not evaluated.'",
+        "gap_closes": [
+          "NIST-800-53-SC-7",
+          "AU-ISM-1546",
+          "NIS2-Art21-patch-management"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-135",
+        "name": "CONTROL-PANEL-PLUGIN-USER-SURFACE-MUST-NOT-REACH-ROOT-OPERATIONS",
+        "description": "The BIG-IP Configuration utility is the constrained user surface this control governs. The packet describes a user authenticated to that utility injecting SQL which the utility passes to its backing database, ending in arbitrary system-command execution — a web administration surface wired straight through to system commands, which is the pattern the control forbids. Applied to this appliance it means the system-command capability sits behind its own authorization boundary that the Configuration utility can only reach through a constrained, parameterized interface, so the utility's own query handling is not the single thing standing between a low-privileged authenticated role and command execution on the device. This is exactly why the least-privilege and identity-and-access gaps cited on this entry can pass their attestations while the flaw stays fully exploitable: role assignment inside the utility is real, auditable and correctly configured, and the injection crosses it — once the SQL reaches the database the role the account holds stops being consulted. Preconditions, both load-bearing: the boundary is a property the vendor's fixed release establishes, so this control states what to verify rather than something the operator implements; and until that release and its reboot land, the only operator-side lever is limiting which accounts can authenticate to the Configuration utility, which bounds the population that can attempt it but closes nothing, because the packet's pairing with CVE-2023-46747 lets an attacker obtain the needed reach while holding no account at all. Because exploitation is confirmed, an appliance whose Configuration utility was reachable during the exposure window needs its administrative accounts and configuration examined against a known-good baseline rather than being closed out on the upgrade.",
+        "evidence": "Packet: CWE-89; 'An authenticated SQL injection vulnerability exists in the BIG-IP Configuration utility which may allow an authenticated attacker with network access to the Configuration utility through the BIG-IP management port and/or self IP addresses to execute arbitrary system commands', and the attack_vector adds that it is in practice paired with CVE-2023-46747 to gain the needed unauthenticated reach. Cited gaps on this entry include NIST-800-53-AC-6 (Least Privilege) and UK-CAF-B2 (Identity and access control). active_exploitation confirmed; poc_available true; CISA KEV-listed 2023-10-31; CVSS 8.8; RWEP 75. patch_available true; live_patch_available false, remediation requires applying the fixed release and rebooting.",
+        "gap_closes": [
+          "NIST-800-53-AC-6",
+          "UK-CAF-B2"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-032",
+        "name": "PERIMETER-COMPROMISE-REBUILD-NOT-PATCH",
+        "description": "The outcome the packet records is arbitrary system-command execution on the BIG-IP itself — the traffic and TLS-termination boundary for whatever sits behind it — so an attacker who reached the Configuration utility read the appliance's configuration and key material before any fixed release was applied. With exploitation confirmed and a PoC public, the default for a unit whose Configuration utility was reachable during the exposure window is configuration capture for forensics, rebuild from vendor media onto a supported release, and rotation of every credential and key the appliance held, rather than upgrading the running unit and closing the ticket. Upgrading in place repairs the injection and preserves everything the attacker altered — and on this appliance that specifically includes the running configuration, which is carried forward through the upgrade, so a rogue administrative account or an altered configuration object survives the remediation that the vulnerability-management record shows as complete. Preconditions and limits: this is the conservative incident-response default and it costs an outage on a device the estate's traffic depends on. The packet gives no per-unit indicator of compromise, so the trigger has to be reachability of the Configuration utility during the window opened by the 2023-10-31 KEV listing, not a confirmed detection. Where an operator can demonstrate the utility was never reachable from any segment outside the administrator network for the whole of that window, the patch-and-reboot path the packet records is sufficient — and demonstrate means evidence from the network boundary, not the assertion that the appliance is internal.",
+        "evidence": "Packet: the injected SQL results in arbitrary system-command execution on the BIG-IP, reachable through the management port and/or self IP addresses, in practice paired with CVE-2023-46747 for unauthenticated reach. active_exploitation confirmed; poc_available true; CISA KEV-listed 2023-10-31; CVSS 8.8; RWEP 75. patch_available true; live_patch_available false with the note 'No vendor live-patch mechanism; remediation requires applying the fixed release and rebooting' — the packet records the fixed release as the remediation, and this control is the statement that on a confirmed-exploited appliance the fixed release alone does not restore trust in the unit.",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.8"
+        ]
+      }
+    ]
   },
   "CVE-2023-46747": {
     "name": "F5 BIG-IP Configuration Utility Authentication Bypass Vulnerability",
@@ -51020,7 +51742,39 @@
         "adequate": false,
         "gap": "Essential-8 rapid-patch targets for internet-facing services are the right control but are commonly unmet for BIG-IP, leaving the preauth path open."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-030",
+        "name": "PERIMETER-DEVICE-ZERODAY-SLA-TIER",
+        "description": "BIG-IP sits on the trust boundary it protects, and the packet has an unauthenticated attacker with network access reaching endpoints that execute arbitrary system commands on it — so the appliance-patch window that treats a load balancer as infrastructure to be touched during a change weekend is the wrong tier for this defect. For this device the requirement is a clock that runs from the 2023-10-31 KEV listing to the moment each BIG-IP has both taken the fixed release and rebooted, because the packet records no vendor live-patch mechanism and states remediation requires applying the fixed release and rebooting. In an HA pair the reboot is the step that gets deferred — the standby takes the update, the failover is scheduled for a later window, and the unit carrying production traffic keeps running vulnerable code — so completion has to be measured per unit as running version plus last restart, not as a change ticket recording that the fixed release was deployed to the cluster. Public PoC and confirmed in-the-wild exploitation are both recorded on this entry, so no deferral rests on exploitation being theoretical. Precondition on the interim lever: restricting where the Configuration utility answers bounds exposure but does not remove it. The packet names the reach paths as the management port and/or self IP addresses, so removing the utility from self IPs cuts one of the two named paths while an attacker positioned on the management network still reaches it and still needs no credential.",
+        "evidence": "Packet entry 'F5 BIG-IP Configuration Utility Authentication Bypass Vulnerability' (CWE-288, CWE-306), cisa_kev true with kev_date 2023-10-31 and active_exploitation 'confirmed'; cvss 9.8, rwep_score 79, poc_available true. Vector: 'Undisclosed requests may bypass configuration utility authentication, allowing an attacker with network access to the BIG-IP system through the management port and/or self IP addresses to execute arbitrary system commands.' patch_available true, live_patch_available false, live_patch_notes: 'No vendor live-patch mechanism; remediation requires applying the fixed release and rebooting.'",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "NIS2-Art21-patch-management"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-128",
+        "name": "APPSERVER-REMOTING-PROTOCOL-RESTRICTION",
+        "description": "The Configuration utility's authentication is enforced at its HTTP front end, and the packet's exploitation path goes around it: a request smuggled to the internal AJP connector reaches endpoints that execute arbitrary system commands without the authentication decision ever being taken. That is the remoting-protocol class this control governs — an internal, non-HTTP protocol endpoint that acts on requests it assumes the fronting layer already authenticated, sitting behind every perimeter and web-tier hardening measure the deployment is audited against. Bound to BIG-IP, the control means the internal connector must make its own authorization decision rather than inheriting a verdict from the layer in front of it, so a request crafted at the HTTP boundary cannot be delivered to it in a form that skips authentication. That property is what the fixed release establishes; the control states what to verify, it does not implement it. The operator-side half is the reach constraint the packet's own text supports: the Configuration utility answers only on the management network, with access via self IP addresses removed. This is also why the identity gaps recorded here are the right ones — the attacker never presents credentials, so no BIG-IP account is abused, no multi-factor prompt is ever raised, and an attestation that every administrator authenticates with unique, strongly-authenticated accounts passes cleanly while the path stays open. Precondition: the self-IP restriction removes one of the two named reach paths and not the other. An attacker with a foothold on the management segment, or on an administrator workstation inside it, satisfies the packet's stated access requirement in full, and the restriction is unavailable in deployments that must expose the utility on a self IP.",
+        "evidence": "Packet attack_vector: 'An unauthenticated attacker smuggles a request to the internal AJP connector, bypassing Configuration-utility authentication and reaching endpoints that execute arbitrary system commands — typically chained with CVE-2023-46748 for the SQLi command primitive.' CWE-288 and CWE-306 are the recorded weaknesses. Vector names the reach paths as 'the management port and/or self IP addresses'. Citing gaps NIST-800-53-SC-7 (Boundary Protection), UK-CAF-B2 (Identity and access control) and NIST-800-53-IA-2 (Identification and Authentication) are recorded against this entry.",
+        "gap_closes": [
+          "NIST-800-53-SC-7",
+          "UK-CAF-B2",
+          "NIST-800-53-IA-2"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-032",
+        "name": "PERIMETER-COMPROMISE-REBUILD-NOT-PATCH",
+        "description": "The packet's outcome is arbitrary system-command execution on the appliance by an unauthenticated attacker, with exploitation confirmed in the wild and a public PoC recorded — the exact conditions under which patch-in-place is the wrong default. Applying the fixed release and rebooting closes the smuggling path; it removes nothing an attacker ran through it beforehand, and on a BIG-IP that means the device's configuration and the credentials and keys stored on it are the exposure, not the vulnerable binary. For any BIG-IP whose Configuration utility was reachable from a position the operator cannot vouch for during the window that opened with the 2023-10-31 KEV listing, the runbook's default should be to treat the unit as attacker-held: capture and review its configuration, rebuild from a known-good baseline rather than upgrading in place, and rotate the credentials and certificates the device held. The packet's chained primitive (CVE-2023-46748) executes system commands through a path where the attacker never authenticates, so there is no failed-login or session artifact for an authentication-centric hunt to find, and a vulnerability-management control that closes on 'fixed release installed' records the device as remediated on exactly the evidence the attack does not produce. Precondition: this is an incident-response default, not a blanket instruction. It applies to units whose utility was reachable from an untrusted position during the exposure window; a unit that can be shown to have been unreachable from any untrusted segment for the whole window is a patch item. Where a rebuild genuinely cannot be absorbed by the service, the fallback is not patch-and-close — it is patch, rotate everything the device held, and carry the unit as an open triage item rather than a closed finding.",
+        "evidence": "Packet records active_exploitation 'confirmed', poc_available true, kev_date 2023-10-31, rwep_score 79. Vector states an attacker with network access can 'execute arbitrary system commands'; attack_vector states the attacker is unauthenticated and that the path is 'typically chained with CVE-2023-46748 for the SQLi command primitive'. live_patch_available false; live_patch_notes: 'No vendor live-patch mechanism; remediation requires applying the fixed release and rebooting.' Citing gap ISO-27001-2022-A.8.8 (Management of technical vulnerabilities) is recorded against this entry.",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.8"
+        ]
+      }
+    ]
   },
   "CVE-2023-5631": {
     "name": "Roundcube Webmail Persistent Cross-Site Scripting (XSS) Vulnerability",
@@ -51762,7 +52516,39 @@
         "adequate": false,
         "gap": "Hardening guidance rarely covers CI/CD platforms; TeamCity admin APIs remained broadly reachable rather than restricted to trusted networks."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-129",
+        "name": "ENTERPRISE-ERP-MANAGEMENT-FUNCTION-AUTHENTICATION",
+        "description": "The packet puts the failure in a REST endpoint on the TeamCity server: an unauthenticated caller reaches it and mints an administrator access token, and from that moment on every action the attacker takes — altering build configuration, executing code on the server — is a properly authorized administrator API call that the product itself issued him the credential for. Bound to this product, the control means each administrative function on the TeamCity server authorizes its caller itself rather than inheriting a verdict from the REST layer that fronts it — token issuance first, then build-configuration and build-step modification — and that the server's REST and web surfaces answer only from segments with an actual build-system role rather than from any network that can route to them. This is why the identity-and-access-control gap is recorded against this entry rather than satisfied by it: the attacker never authenticates as any TeamCity user, so per-account privilege scoping and role assignment are never consulted, and the account model an identity attestation examines is bypassed rather than abused. The user-application-hardening gap misses it for a different reason — that control governs settings on workstation software and never examines a build server's API surface at all. Distinguishing test: from a segment with no developer, build-agent or integration role, issue unauthenticated requests to the TeamCity REST endpoints that create access tokens and modify build configuration on a staging server, and confirm each is refused before the function runs; an attestation that every TeamCity administrator authenticates at login and holds a minimally-scoped role passes cleanly while this path stays open. Precondition, and it is the half most often over-claimed: the endpoint-side authorization is a property the vendor fixed release establishes — this control states what to verify, it does not implement it. Until the server is on a release at or above 2023.05.4, restricting which segments can reach the REST surface bounds the population that can send the request but leaves the endpoint fully exploitable to anything inside a permitted segment, and it is simply unavailable where that surface must stay reachable, since serving developers, build agents and integrations is what a TeamCity server exists to do.",
+        "evidence": "Packet attack_vector: 'An unauthenticated attacker abuses a REST-endpoint authentication bypass to mint an administrator access token, then uses privileged API access to alter build configuration and execute arbitrary code on the TeamCity server, pivoting into the software supply chain.' CWE-288 and CWE-306 (missing authentication for critical function). Packet vector names the fix boundary: 'In JetBrains TeamCity before 2023.05.4 authentication bypass leading to RCE on TeamCity Server was possible.' CISA KEV-listed 2023-10-04, active_exploitation confirmed, poc_available true, RWEP 72 against CVSS 9.8. Citing gaps on this entry include NIST-800-53-SC-7 (Boundary Protection), UK-CAF-B2 (Identity and access control) and AU-Essential-8-App-Hardening (User application hardening).",
+        "gap_closes": [
+          "UK-CAF-B2",
+          "NIST-800-53-SC-7",
+          "AU-Essential-8-App-Hardening"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-078",
+        "name": "ENDPOINT-MGMT-DEPLOYMENT-CHANNEL-INTEGRITY",
+        "description": "The primitive here does not end at code execution on one server: the packet has the attacker altering build configuration and pivoting into the software supply chain, so the exposed asset is every artifact that TeamCity builds and every credential it holds for the registries and deployment targets downstream of it. Treat the build configurations, build steps and artifact output directories as a privileged distribution channel rather than as application data — file-integrity-monitor them, alert on any change that does not correspond to a sanctioned commit or an operator change record, and inventory the access tokens the server has issued so that a token minted through the bypass can be found and revoked rather than surviving as a valid administrator credential. This is the control that still has value after the fixed release lands, because the upgrade closes the authentication bypass and removes nothing that was already written through it: an injected build step, an attacker-created administrator token, and an artifact already published to a downstream consumer all survive the upgrade untouched. None of them will surface in authentication logging either, because from the moment the token was minted every request was properly authorized, and no signature-based tooling has anything to match against a build step that is simply an extra command in a legitimate pipeline. Distinguishing test: introduce an out-of-band change to a build step on a staging TeamCity server and confirm it raises an alert with no corresponding change record. Precondition: this is detection and containment, not remediation — it does not close the bypass, and it only reaches configuration and artifact paths that were being baselined before the exposure window opened; a baseline captured after the fact records the attacker's state as normal. An estate that closes this CVE when every server reports 2023.05.4 or later has established that the door is shut, not that nobody came through it during the window that the packet's confirmed in-the-wild exploitation implies.",
+        "evidence": "Packet attack_vector states the outcome reaches beyond the host: the attacker 'uses privileged API access to alter build configuration and execute arbitrary code on the TeamCity server, pivoting into the software supply chain.' active_exploitation confirmed and poc_available true, so exploitation during the exposure window is the packet's stated condition rather than a hypothetical. live_patch_available false with live_patch_notes 'No vendor live-patch mechanism; remediation requires applying the vendor fixed release', so the only recorded remediation is a version change, which acts on the code path and not on state an attacker has already written. Citing gap ISO-27001-2022-A.8.8 (Management of technical vulnerabilities).",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.8"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-001",
+        "name": "CISA-KEV-RESPONSE-SLA",
+        "description": "For this CVE the KEV clock opened 2023-10-04 and the packet records exactly one remediation: moving the TeamCity server to the vendor fixed release, which its own vector places at 2023.05.4 or later. There is no live-patch mechanism registered and the packet names no vendor mitigation rule, so nothing else exists to deploy inside the window — the upgrade is not the preferred option among several, it is the only one. Completion therefore has to be measured as each TeamCity server reporting a version at or above 2023.05.4, not as the upgrade being approved, scheduled or downloaded, because a build server is precisely the asset whose maintenance window gets deferred to avoid interrupting in-flight pipelines, and that deferral recorded as 'remediation in progress' is how this specific flaw stays live. The packet's numbers argue against the deferral: a public PoC, confirmed in-the-wild exploitation, and an RWEP of 72 against a CVSS of 9.8 describe a flaw whose exploitation cost is already paid by someone else. Priority ordering follows reachability rather than server size, since the packet makes an unauthenticated network request the only precondition — the attacker holds no credential, so any TeamCity server a hostile caller can route to is in scope and a lightly-used one is no safer than the flagship instance. Precondition on the interim: until a server takes the upgrade, restricting who can reach its REST surface bounds the attempt but does not close it, and offers nothing on a server that must remain reachable to its agents and developers.",
+        "evidence": "Packet: cisa_kev true, kev_date 2023-10-04, active_exploitation confirmed, poc_available true, rwep_score 72, cvss 9.8. patch_available true; live_patch_available false with live_patch_notes 'No vendor live-patch mechanism; remediation requires applying the vendor fixed release.' Fixed-version boundary from the packet vector: 'In JetBrains TeamCity before 2023.05.4.' Packet attack_vector establishes the no-credential precondition: 'An unauthenticated attacker abuses a REST-endpoint authentication bypass...' Citing gaps NIST-800-53-SI-2 (Flaw Remediation) and NIS2-Art21-vulnerability-handling.",
+        "gap_closes": [
+          "NIST-800-53-SI-2",
+          "NIS2-Art21-vulnerability-handling"
+        ]
+      }
+    ]
   },
   "CVE-2023-28229": {
     "name": "Microsoft Windows CNG Key Isolation Service Privilege Escalation Vulnerability (CVE-2023-28229)",
@@ -51819,7 +52605,23 @@
         "adequate": false,
         "gap": "Patch-application timeframes for endpoint OS updates commonly exceed the window in which a public PoC enables reliable local escalation."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-145",
+        "name": "ACTIVELY-EXPLOITED-OS-LPE-REMEDIATION-WINDOW-ENFORCEMENT",
+        "description": "The packet's path is a low-privileged local process making racing concurrent RPC calls to the CNG Key Isolation service inside lsass.exe, where flawed critical-section handling of reference counters produces a use-after-free that yields limited SYSTEM privileges. Everything the attacker needs is the ability to run code as an ordinary user on the machine, so the population to drive first is the one where that is the normal operating state rather than an anomaly — multi-user session hosts, shared and kiosk workstations, jump hosts, any machine where non-administrative interactive users are expected. For this CVE the control means the Windows update carrying the fix is driven across that population on the clock that opened with the 2023-10-04 KEV listing rather than folded into the next monthly rollup, with completion measured per host as the running build against the fixed build for its SKU rather than as 'approved', 'downloaded' or 'installed' in the management console. The packet records no live-patch mechanism and states remediation requires applying the fixed release and rebooting, so a host that has taken the update and deferred the restart is still running the vulnerable lsass.exe and must be counted exposed — and on a shared session host that restart is the step most likely to be deferred, because taking it evicts logged-on users, which is exactly how this remediation is misreported as complete. The control's second half is the load-bearing one here: the attacker is already an authorized low-privilege user, so tightening account privilege does not contain the escalation, which is why the least-privilege gap cited on this entry can attest cleanly while the flaw stays fully exploitable. Precondition, stated plainly because there is no compensating control to offer in its place: before the fixed build lands there is no operator-side lever that removes this path. The packet's only access requirement is a low-privileged local process, which every interactive user of the host satisfies; restricting who may log on bounds the population that can attempt the race but does not close it for anyone holding a legitimate session, and nothing in the packet describes a configuration, feature toggle or service state that takes the CNG Key Isolation RPC interface out of reach. Priority follows the packet rather than the 7.0 CVSS: a public PoC, confirmed in-the-wild exploitation and RWEP 75 make this the second stage of a chain — the step that converts a foothold into SYSTEM — not a standalone endpoint item to be deprioritized behind remote CVEs.",
+        "evidence": "Packet attack vector: 'A low-privileged local process makes racing concurrent RPC calls to the CNG Key Isolation service in lsass.exe; flawed critical-section handling of reference counters produces a use-after-free the attacker leverages to gain limited SYSTEM privileges.' CWE-591. CISA KEV-listed 2023-10-04, active_exploitation 'confirmed', poc_available true, CVSS 7.0, RWEP 75. patch_available true; live_patch_available false, with the packet's note stating there is no vendor live-patch mechanism and that remediation requires applying the fixed release and rebooting.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "ISO-27001-2022-A.8.8",
+          "NIS2-Art21-patch-management",
+          "NIST-800-53-SI-2",
+          "NIST-800-53-AC-6",
+          "UK-CAF-B4"
+        ]
+      }
+    ]
   },
   "CVE-2023-4211": {
     "name": "Arm Mali GPU Kernel Driver Use-After-Free Vulnerability",
@@ -53286,7 +54088,30 @@
         "adequate": false,
         "gap": "Periodic technical-vulnerability management does not compel the input-validation review of the WP_Query surface that would prevent untrusted data reaching author__not_in in the first place."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-085",
+        "name": "DB-ABSTRACTION-LAYER-PARAMETERIZATION-VERIFICATION",
+        "description": "WP_Query is WordPress core's query builder, and the packet places the defect inside it: core mishandles the author__not_in parameter, so SQL injection follows whenever untrusted input reaches that parameter. That is the exact shape this control governs — parameterization has to be verified where the query is built, not inferred from a plugin's input filtering or from a perimeter WAF sitting in front of the site, because every caller of WP_Query inherits whatever the builder does with the parameter. The reachability detail in the packet is what makes the verification non-obvious and is the reason a review can wrongly clear this: the parameter is reachable unauthenticated only when chained with CVE-2026-63030's REST batch-route desynchronization, so an assessment that exercises only authenticated code paths concludes the parameter is not attacker-reachable and the injection unexploitable, while the wp2shell chain reaches it on a default install. Distinguishing test: on a staging install, pass SQL metacharacters through author__not_in — both directly and through the REST route the chain desynchronizes — and confirm the query builder parameterizes rather than concatenates. Precondition: the parameterization repair is what the vendor releases ship; this control is the standing verification requirement and gives nothing on its own to a site still on a pre-fix build, and it does not address a site already exploited through the chain.",
+        "evidence": "Packet vector: 'WordPress Core mishandles the author__not_in parameter of WP_Query, allowing SQL injection when untrusted input reaches the parameter; chained with CVE-2026-63030 for unauthenticated remote code execution.' Attack vector: reachable unauthenticated 'only when chained with CVE-2026-63030's REST batch-route desynchronization', forming 'the data-access half of the wp2shell pre-auth RCE chain against default WordPress installs'. CWE-89; poc_available true; active_exploitation confirmed. Citing gap UK-CAF-B4 (System security).",
+        "gap_closes": [
+          "UK-CAF-B4"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-001",
+        "name": "CISA-KEV-RESPONSE-SLA",
+        "description": "This entry is the case a severity-banded patch policy handles worst. CVSS 5.9 routes it into a low or medium queue with a 30-to-90-day window under the flaw-remediation and technical-vulnerability controls cited here, while the packet records CISA KEV listing on 2026-07-21, confirmed in-the-wild exploitation, a public PoC, RWEP 78, and a role as the data-access half of a pre-auth RCE chain against default installs. The control's requirement bound to this CVE is that the clock runs from the KEV listing and the target is the core version the packet names — 6.8.6, 6.9.5 or 7.0.2 — with the CVSS band explicitly not used to defer. The packet records WordPress.org enabling forced automatic security updates, which shortens the window wherever that channel is active, but a delivery mechanism is not a completion record: the evidence an operator can defend is the core version each site actually reports, per site, not the assumption that a push landed everywhere. That distinction is the whole gap here, because a vulnerability-management programme that closes this item on 'WordPress pushed the fix' has recorded a vendor action rather than a state of its own estate. Precondition: the packet records no live-patch mechanism, so the upgrade itself is the only remediation; and because exploitation is confirmed and the chain ends in remote code execution, a site that was internet-reachable between the listing and its own upgrade needs to be examined rather than closed on the version number.",
+        "evidence": "CISA KEV listed 2026-07-21; active_exploitation confirmed; poc_available true; RWEP 78 against CVSS 5.9; patch_available true; live_patch_available false; live_patch_notes: 'No live-patch mechanism; WordPress.org enabled forced automatic security updates. Remediation is upgrading to 6.8.6 / 6.9.5 / 7.0.2.' Citing gaps AU-Essential-8-Patch, ISO-27001-2022-A.8.8, NIST-800-53-SI-2, NIS2-Art21-vulnerability-management.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "ISO-27001-2022-A.8.8",
+          "NIST-800-53-SI-2",
+          "NIS2-Art21-vulnerability-management"
+        ]
+      }
+    ]
   },
   "CVE-2026-63030": {
     "name": "WordPress Core Interpretation Conflict Vulnerability",
@@ -56114,7 +56939,38 @@
         "adequate": false,
         "gap": "A.8.8 would rank CVE-2023-32439 critical once KEV-listed, yet technical-vulnerability management is reactive here — the exploit predates the advisory, so only prompt Rapid Security Response deployment closes the WebKit RCE."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-056",
+        "name": "MOBILE-ENDPOINT-MDM-ENFORCED-KEV-SLA",
+        "description": "The packet's fixed builds span two iOS majors and two platforms at once — iOS 16.5.1 and iPadOS 16.5.1, iOS 15.7.7 and iPadOS 15.7.7, macOS Ventura 13.4.1, and Safari 16.5.1 — so 'the estate is current' is not a statement about this CVE. A device held on the 15.x branch needs 15.7.7 specifically, and a Mac not moving to Ventura 13.4.1 needs the standalone Safari 16.5.1 as its own item. Applied here, the control means the management platform pushes and enforces each of those builds against the population it applies to, on a clock tied to the 2023-06-23 KEV listing rather than a monthly ring, with user deferral disallowed — because the packet records the fix as arriving in an OS/Safari update (or Rapid Security Response) that requires a device restart to apply, and the restart is the step a user postpones indefinitely on a phone they are carrying. Completion is measured as the restarted build the device reports, not as a profile marked installed or an update marked downloaded. This cannot be folded into an ordinary application-update policy for one browser: the defect is in WebKit, and the packet's attack vector names Safari or any WebKit-based HTML parser as the trigger, so on these platforms every app that renders web content is a delivery path and the OS update is the only action that reaches all of them together. Precondition: this reaches only enrolled devices. A personally-owned or unenrolled iPhone, iPad or Mac rendering the same WebKit content is untouched by the policy, and for that population this control delivers nothing — closing that half needs an access condition, not an update push.",
+        "evidence": "Packet entry 'Apple Multiple Products WebKit Type Confusion Vulnerability (CVE-2023-32439)' (CWE-843), cisa_kev true with kev_date 2023-06-23 and active_exploitation 'confirmed'; cvss 8.8, rwep_score 52, poc_available false. Vector: 'This issue is fixed in iOS 16.5.1 and iPadOS 16.5.1, iOS 15.7.7 and iPadOS 15.7.7, macOS Ventura 13.4.1, Safari 16.5.1. Processing maliciously crafted web content may lead to arbitrary code execution. Apple is aware of a report that this issue may have been actively exploited.' attack_vector names 'Safari (or any WebKit-based HTML parser)'. patch_available true, live_patch_available false, live_patch_notes: 'No live-patch mechanism; Apple ships the fix in an OS/Safari update (or Rapid Security Response) that requires a device restart to apply.'",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "NIS2-Art21-patch-management",
+          "NIST-800-53-SI-2"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-126",
+        "name": "MOBILE-OS-MINIMUM-PATCH-LEVEL-ENFORCEMENT",
+        "description": "The population that stays exposed after the push is the one the push cannot reach or the user keeps deferring, and this packet's fix list makes that population concrete: devices sitting on the iOS/iPadOS 15.7.7 branch are typically there because they cannot take 16.5.1, and macOS hosts that will not move to Ventura 13.4.1 need Safari 16.5.1 tracked separately. The requirement here is that the fixed build works as an access condition rather than a dashboard row — a device below iOS/iPadOS 16.5.1 or 15.7.7, or a Mac below macOS Ventura 13.4.1 that has not taken Safari 16.5.1, is denied mail, VPN and document access until it reports at or above the fix. The distinguishing test is to enrol a device pinned below those builds and confirm the policy actually refuses it access to protected resources; an estate that surfaces the stale build on a compliance report while the device keeps its mailbox has recorded the exposure rather than removed it. Because the packet records the fix as requiring a device restart, a device that has downloaded the update but not restarted must be treated by the policy as below the build, not above it. Scope it to what the packet names — iOS, iPadOS, macOS and Safari. The fixed-version list is Apple's, and the packet gives no basis for treating non-Apple browsers in the estate as instances of this CVE. Precondition: an access condition bounds what a device can reach next, it does not protect the device itself. The trigger is processing maliciously crafted web content leading to arbitrary code execution in the renderer, so a device that already rendered the attacker's page is compromised whether or not it holds corporate data, and it belongs on the incident path rather than the enrolment path.",
+        "evidence": "Packet vector lists the fixed builds as iOS 16.5.1 and iPadOS 16.5.1, iOS 15.7.7 and iPadOS 15.7.7, macOS Ventura 13.4.1, Safari 16.5.1, and states 'Processing maliciously crafted web content may lead to arbitrary code execution.' live_patch_available false with live_patch_notes recording that the fix ships in an OS/Safari update (or Rapid Security Response) 'that requires a device restart to apply'. active_exploitation 'confirmed'; citing gap ISO-27001-2022-A.8.8 (Management of technical vulnerabilities) is recorded against this entry.",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.8"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-121",
+        "name": "MOBILE-ZERO-CLICK-HARDENING",
+        "description": "The packet records confirmed exploitation with no public PoC, and describes the bug as a stage in a browser-based exploit chain — code execution in the renderer that a following stage has to escalate. That combination points at limited, targeted use rather than commodity mass exploitation, and for the cohort plausibly inside that targeting set the reboot-gated OS update is not fast enough on its own: the window between the 2023-06-23 KEV listing and a completed restart across the fleet is exactly when such a chain is worth spending. The requirement is that those users are already in the platform's reduced-attack-surface mode — untrusted web content not taken through the full WebKit path, message-borne content and link previews not rendered automatically — assigned before the next disclosure rather than switched on in response to this one, because the mode only helps if it was already active when the page was served. Precondition, and it is why this is never a substitute for the update: the mode narrows the delivery path into the WebKit parser, it does not repair the type confusion, and it does nothing for a device that has already processed the attacker's content. It applies only to the cohort it is assigned to and only where the platform offers it, so the rest of the estate stays on the update-and-restart path, which the packet records as the only remediation with no live-patch mechanism available.",
+        "evidence": "Packet attack_vector: 'A type-confusion bug in WebKit is triggered when Safari (or any WebKit-based HTML parser) processes maliciously crafted web content, corrupting object typing to achieve arbitrary code execution in the renderer as a stage in a browser-based exploit chain.' active_exploitation 'confirmed' with poc_available false; kev_date 2023-06-23. Vector: 'Apple is aware of a report that this issue may have been actively exploited.' live_patch_available false; live_patch_notes: 'No live-patch mechanism; Apple ships the fix in an OS/Safari update (or Rapid Security Response) that requires a device restart to apply.' Citing gap UK-CAF-B4 (System security) is recorded against this entry.",
+        "gap_closes": [
+          "UK-CAF-B4"
+        ]
+      }
+    ]
   },
   "CVE-2023-20867": {
     "name": "VMware Tools Authentication Bypass Vulnerability",
@@ -56663,7 +57519,43 @@
         "adequate": false,
         "gap": "A.8.8 technical-vulnerability management often under-inventories network appliances versus servers, so a vulnerable FortiGate can miss the remediation cycle even after FG-IR-23-097 and KEV listing."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-030",
+        "name": "PERIMETER-DEVICE-ZERODAY-SLA-TIER",
+        "description": "The packet puts the defect in the FortiOS/FortiProxy SSL-VPN web interface and reaches it with a crafted pre-authentication request that overflows a heap buffer for code execution on the appliance with no credentials — the device terminating remote access is itself the vulnerable software, which is why an ordinary network-device patch window is the wrong tier for it. Bound to this CVE, the tier means the clock runs from the 2023-06-13 KEV listing to the completed reboot of every FortiGate and FortiProxy unit whose running build falls inside the packet's affected list (FortiOS 7.2.4 and below, 7.0.11 and below, 6.4.12 and below, 6.0.16 and below; FortiProxy 7.2.3 and below, 7.0.9 and below, 2.0.12 and below, and 1.2 and 1.1 in all versions), with the only permitted alternative being isolation of the SSL-VPN interface rather than a later change window. Two packet facts make the standard window unusable here. First, the fix is a build upgrade that reboots the appliance and there is no live-patch mechanism, so a unit with the image staged but the reboot deferred is still running the vulnerable code and must be counted exposed — on an edge appliance that reboot is precisely the step that gets deferred, because taking it drops every live VPN session, and a deferral recorded as 'patched' is the specific way this remediation goes wrong. Second, FortiProxy 1.2 and 1.1 are listed as affected in all versions, so for a unit on either train there is no fixed build to reach inside the train and remediation is a move to a train that has one — a distinction any compliance check that only asks 'is the build newer than the one installed' will not make. The distinguishing test: produce, per unit, the running build and the time of its last restart and show the build is outside the packet's affected list; a fleet report showing an approved or downloaded upgrade against a unit that has not restarted records an intention, not a remediation. Precondition on the isolation alternative: interface isolation is available only on a unit whose SSL-VPN is not carrying production remote access; where it is the organization's remote-access path, that lever does not exist and the reboot-bearing upgrade is the only remediation, taken early rather than at the next window. Priority follows the packet rather than the CVSS band — pre-authentication reach, a public PoC, and confirmed in-the-wild exploitation.",
+        "evidence": "Packet: CISA KEV-listed 2023-06-13, active_exploitation 'confirmed', poc_available true, CVSS 9.8, RWEP 79. Attack vector: a crafted pre-authentication request to the FortiOS/FortiProxy SSL-VPN web interface overflows a heap buffer (CWE-122, CWE-787), giving remote code execution on the appliance with no credentials. patch_available true; live_patch_available false, with the packet's live-patch note stating there is no live-patch mechanism for FortiOS and that remediation requires upgrading to a fixed FortiOS/FortiProxy build, which reboots the appliance. Affected list per the packet: FortiOS 7.2.4 and below, 7.0.11 and below, 6.4.12 and below, 6.0.16 and below; FortiProxy 7.2.3 and below, 7.0.9 and below, 2.0.12 and below, 1.2 all versions, 1.1 all versions.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "ISO-27001-2022-A.8.8",
+          "NIS2-Art21-patch-management",
+          "NIST-800-53-SI-2"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-025",
+        "name": "CONFIGURATION-SIDE-LIVE-PATCH-INVENTORY",
+        "description": "This entry has a configuration-side mitigation path that does not wait on the reboot-bearing upgrade — the packet names disabling SSL-VPN as the interim mitigation — and the control's requirement is that the path be inventoried and rehearsed before the disclosure rather than discovered during it. Applied to this fleet, that means knowing per unit whether SSL-VPN is carrying production remote access or is enabled-but-unused, because those two populations have entirely different options: the second can be taken out of reach in minutes with a configuration change and no restart, while the first cannot be touched without removing remote access and must go straight to the upgrade. An estate that cannot answer that question per unit spends the exposure window scheduling a maintenance slot for units that never needed one. The distinguishing test is behavioural, not a configuration read: after applying the change on a unit, confirm the SSL-VPN web interface no longer answers on the interfaces that previously served it — a checkbox in a saved configuration is not evidence that the pre-authentication listener is gone. Preconditions, both load-bearing and neither optional to state. This lever exists only where the SSL-VPN service is not the operational remote-access path; the packet's own framing is that it is an interim mitigation, not the remediation, and the remediation remains the fixed build and its reboot. And because the packet records exploitation confirmed in the wild from the 2023-06-13 KEV listing with a public PoC, disabling the service on a unit that was already exploited closes the entry path without removing anything the attacker left behind — such a unit belongs on the incident path, not the configuration path.",
+        "evidence": "Packet live-patch note, verbatim: 'No live-patch mechanism for FortiOS; remediation requires upgrading to a fixed FortiOS/FortiProxy build, which reboots the appliance. Disabling SSL-VPN removes the attack surface as an interim mitigation.' patch_available true, live_patch_available false. The packet locates the flaw in the SSL-VPN interface of FortiOS/FortiProxy and describes the triggering request as pre-authentication. CISA KEV-listed 2023-06-13, active_exploitation 'confirmed', poc_available true.",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.8",
+          "NIS2-Art21-patch-management",
+          "UK-CAF-B4"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-032",
+        "name": "PERIMETER-COMPROMISE-REBUILD-NOT-PATCH",
+        "description": "The packet gives this entry both halves of the condition this control exists for: a pre-authentication remote-code-execution path on the appliance itself, and exploitation confirmed in the wild with a public PoC. The upgrade closes the heap overflow; it removes nothing an attacker placed on a unit that was exploited before the upgrade landed. So for any FortiGate or FortiProxy whose SSL-VPN interface was reachable during the exposure window that opened at the 2023-06-13 KEV listing, the default response is configuration extraction, rebuild from a known-good baseline on a fixed build, and rotation of the credentials the unit held — not patch-in-place followed by closing the ticket. Precondition, and it is the half most often skipped: this presumes the operator can determine whether a given unit was exploited, and the packet's own attack path makes that hard by construction — the request arrives before any credential check, so there is no failed or anomalous authentication event to find, and a unit whose only record of the window is its own local log cannot answer the question at all, because a successful exploit runs on the device holding that log. Where the evidence to exclude compromise does not exist, the default has to be rebuild rather than 'no evidence of compromise found'. The precondition also bounds the control honestly: rebuilding restores the device, it does not undo use already made of anything the unit stored, which is why the credential rotation is part of the same action rather than a follow-up item.",
+        "evidence": "Packet: active_exploitation 'confirmed', poc_available true, CISA KEV-listed 2023-06-13, RWEP 79, CVSS 9.8. Attack vector: a crafted pre-authentication request to the SSL-VPN web interface overflowing a heap buffer to gain remote code execution on the appliance with no credentials. patch_available true, live_patch_available false, with the packet stating remediation requires upgrading to a fixed FortiOS/FortiProxy build, which reboots the appliance.",
+        "gap_closes": [
+          "NIST-800-53-SI-2",
+          "ISO-27001-2022-A.8.8",
+          "UK-CAF-B4"
+        ]
+      }
+    ]
   },
   "CVE-2023-3079": {
     "name": "Google Chromium V8 Type Confusion Vulnerability (CVE-2023-3079)",
@@ -57370,7 +58262,29 @@
         "adequate": false,
         "gap": "Network segregation (out-of-band management, VTY source ACLs) is the real mitigation; an organization relying on A.8.8-style patch tracking for a 2004 advisory on legacy IOS would not close the reachable Telnet management plane."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-001",
+        "name": "CISA-KEV-RESPONSE-SLA",
+        "description": "The packet gives two remediations for this router flaw with very different clocks, and the SLA has to name both. The fixed IOS train is the closure, but the packet records no live-patch path and a device reload to apply it, so the reload -- not the image copy -- is what the clock runs to; a device with the new image staged but not reloaded still allocates VTY lines the old way when the crafted TCP connections arrive and must be counted as exposed. The second remediation the packet names, disabling or ACL-restricting Telnet/VTY, is the one an operator can actually deploy inside a KEV-listing window on a router whose reload needs a change window, so the SLA is met by recording that compensating control as active with the fixed train dated to the next reload -- not by letting the interim state stand indefinitely as 'mitigated'. The deployment path itself is what this CVE breaks, which is why it cannot be run as a routine patch item: the packet's stated outcome is that the device refuses further Telnet, SSH, RSH and in some cases HTTP management access, so a device already under the attack cannot be reached in-band to receive either the ACL or the new image. The mitigation has to be pre-staged and the out-of-band console path proven working before it is needed, rather than pushed reactively once the lines are consumed. Distinguishing test: per device, produce the running image, the VTY transport and access-class configuration, and evidence that the reload carrying the fixed train actually happened -- a patch report that reads 'current' off a scheduled image push marks a router compliant while it is still executing the pre-reload code.",
+        "evidence": "Packet: cisa_kev true, kev_date 2023-05-19, active_exploitation 'confirmed'. patch_available true, live_patch_available false, live_patch_notes: 'No live-patch for classic Cisco IOS; remediation is upgrading to a fixed IOS train (or disabling/ACL-restricting Telnet/VTY), which requires a device reload.' attack_vector: after all VTY lines are consumed 'the device refuses further Telnet, SSH, RSH and (in some cases) HTTP management access' -- the in-band path an operator would otherwise use to deploy either mitigation. rwep_score 42 against cvss 5.9 with poc_available false: the KEV listing and confirmed exploitation, not the severity band, are what put this on a clock.",
+        "gap_closes": [
+          "AU-Essential-8-Patch"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-128",
+        "name": "APPSERVER-REMOTING-PROTOCOL-RESTRICTION",
+        "description": "The listener this control governs on a Cisco IOS device is the VTY line pool reached through the Telnet and reverse-Telnet ports: a non-HTTP endpoint that answers a TCP connection and allocates a line before any credential is presented, which is why an unauthenticated attacker can exhaust it and why perimeter and web-tier hardening never touch the path. Applied to the affected devices, the requirement is that those ports answer only from the management and jump networks that legitimately drive them -- enforced by an access class on the VTY lines and by upstream ACLs, not assumed from 'the router is internal' -- and that Telnet be removed where the device is administered over SSH, since the packet places the entry point on the Telnet/reverse-Telnet port specifically. Two preconditions have to be stated rather than glossed. First, restricting the source ranges bounds who can open the crafted connections; it does not repair the VTY allocation the flaw abuses, so any host inside the permitted management segment -- a compromised jump host, a contractor laptop -- can still consume every line and lock the operators out of the device. Second, on a unit doing terminal-server/console-server duty the reverse-Telnet listener has to stay reachable from the segments it serves, so for those units restriction narrows the source set and nothing more; there is no configuration that removes the path while the function is in use. Scope this to devices running the affected IOS with those listeners enabled, not to network equipment generally. Distinguishing test: from a general user VLAN, open a TCP connection to the device's Telnet and reverse-Telnet ports and confirm it is dropped before a VTY line is allocated -- a configuration review that shows SSH is enabled and Telnet 'not used' passes cleanly while the port still answers and still consumes a line.",
+        "evidence": "Packet vector: 'Cisco IOS 12.2(15) and earlier allows remote attackers to cause a denial of service (refused VTY (virtual terminal) connections), via a crafted TCP connection to the Telnet or reverse Telnet port.' attack_vector: 'An unauthenticated attacker opens a series of crafted TCP connections to the device's Telnet/reverse-Telnet port until all VTY lines are consumed.' live_patch_notes names 'disabling/ACL-restricting Telnet/VTY' as the alternative to the fixed IOS train. Citing gaps on this entry include ISO-27001-2022-A.8.22 (Segregation of networks), NIS2-Art21-network-security and UK-CAF-B4 (System security). cwe_refs CWE-400.",
+        "gap_closes": [
+          "ISO-27001-2022-A.8.22",
+          "NIS2-Art21-network-security",
+          "UK-CAF-B4"
+        ]
+      }
+    ]
   },
   "CVE-2016-6415": {
     "name": "Cisco IOS, IOS XR, and IOS XE IKEv1 Information Disclosure Vulnerability",
@@ -58425,7 +59339,30 @@
         "adequate": false,
         "gap": "A.5.15 access-control policy is defeated at the root: MinIO discloses the root credential itself pre-auth, so any access-control matrix built on that secret is bypassed the moment the bootstrap endpoint is queried."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-001",
+        "name": "CISA-KEV-RESPONSE-SLA",
+        "description": "MinIO's remediation has three parts and only one of them is a patch, which is why a version-only SLA closes this ticket while the cluster stays owned. The packet's fixed release stops /minio/bootstrap/v1/verify returning process environment variables, but MINIO_ROOT_PASSWORD and MINIO_SECRET_KEY left the process before the upgrade and remain valid after it -- an attacker who read them authenticates as admin against the patched build exactly as they did against the vulnerable one. For this deployment the mitigation the SLA measures is all three of the packet's steps: the upgrade to RELEASE.2023-03-20T20-16-18Z or later, rotation of the exposed MINIO_ROOT_PASSWORD and MINIO_SECRET_KEY, and the service restart that puts the rotated values into the running process. The packet records no live-patch path, so that restart is unavoidable and every node takes it. Scope is the whole cluster rather than a representative node: the packet states the flaw affects cluster deployments and that all users of distributed deployment are impacted, and the disclosure endpoint answers per node, so a rolling upgrade that leaves one node on an affected release leaves the disclosure reachable and the rotation pointless. Distinguishing test: after the upgrade, POST to /minio/bootstrap/v1/verify on each node and confirm no environment content is returned, and separately confirm the pre-incident root credential no longer authenticates -- a flaw-remediation attestation that records the release string and stops there marks the cluster compliant while the leaked credential still works.",
+        "evidence": "Packet: cisa_kev true, kev_date 2023-04-21, active_exploitation 'confirmed', poc_available true, rwep_score 66, cvss 7.5. patch_available true, live_patch_available false, live_patch_notes: 'No vendor live-patch; remediation requires upgrading to MinIO RELEASE.2023-03-20T20-16-18Z or later, rotating the exposed MINIO_ROOT_PASSWORD / MINIO_SECRET_KEY, and restarting the service.' vector: 'In a cluster deployment starting with RELEASE.2019-12-17T23-16-33Z and prior to RELEASE.2023-03-20T20-16-18Z, MinIO returns all environment variables, including MINIO_SECRET_KEY and MINIO_ROOT_PASSWORD... All users of distributed deployment are impacted.' attack_vector places the disclosure at an unauthenticated POST to /minio/bootstrap/v1/verify on a distributed MinIO node.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "NIST-800-53-SI-2",
+          "NIS2-Art21-vulnerability-management"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-078",
+        "name": "ENDPOINT-MGMT-DEPLOYMENT-CHANNEL-INTEGRITY",
+        "description": "The packet's chain does not end at disclosure. With MINIO_ROOT_PASSWORD in hand the attacker authenticates as admin and uses mc admin update to push a malicious binary, so MinIO's administrative API is a code-distribution channel for the server binary itself and a single disclosed credential is the whole gate on it. Treat it as that rather than as storage administration: file-integrity-monitor the MinIO binary and the service configuration on every node of the distributed deployment, alert on any replacement that does not correspond to a sanctioned operator action or a vendor upgrade, and restrict which identities may invoke the administrative update path at all. This is the half that still has value after the fixed release lands, because the upgrade closes the environment-variable disclosure but removes nothing that was pushed through the update path while the credential was exposed -- a node reachable during the exposure window has to be compared against a known-good binary rather than closed on its version string. It is also precisely what the access-control gaps recorded against this entry miss: an attestation that MinIO's admin accounts are controlled examines who is issued the root credential, not that holding it confers replacement of the running server, and rotating the credential afterwards does not evict a binary already installed. Precondition: integrity monitoring detects the substitution, it does not prevent it, and it is worth nothing on a node whose baseline was taken after the deployment was already exposed -- there the attacker's binary is what gets recorded as normal, and the comparison has to be against vendor-published artifacts rather than against the node's own history.",
+        "evidence": "Packet attack_vector: the disclosed MINIO_ROOT_PASSWORD and MINIO_SECRET_KEY are what 'the attacker uses to authenticate as admin and (via mc admin update) push a malicious binary for code execution.' live_patch_notes records remediation as the upgrade plus rotation of MINIO_ROOT_PASSWORD / MINIO_SECRET_KEY plus a service restart -- none of which reverses a binary already pushed. active_exploitation 'confirmed' with poc_available true and kev_date 2023-04-21. Citing gaps include UK-CAF-B2 (Identity and access control) and ISO-27001-2022-A.5.15 (Access control).",
+        "gap_closes": [
+          "UK-CAF-B2",
+          "ISO-27001-2022-A.5.15"
+        ]
+      }
+    ]
   },
   "CVE-2023-27350": {
     "name": "PaperCut MF/NG Improper Access Control Vulnerability",
@@ -58981,7 +59918,30 @@
         "adequate": false,
         "gap": "A.8.8 technical-vulnerability management may rank a local 7.8 EoP below internet-facing criticals, but its confirmed ransomware use as the privilege-escalation link means the CVSS score understated real risk and only prompt OS patching removed the SYSTEM path."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-145",
+        "name": "ACTIVELY-EXPLOITED-OS-LPE-REMEDIATION-WINDOW-ENFORCEMENT",
+        "description": "The packet places this in clfs.sys, the Windows Common Log File System driver, and describes a local attacker crafting a malicious CLFS base log (.BLF) to drive a heap overflow and out-of-bounds write in kernel memory, taking a low-privileged process to SYSTEM. For this CVE the control means the April 2023 Windows cumulative security update is driven across the affected estate on the KEV clock that opened 2023-04-11 rather than folded into the next monthly maintenance cycle, with completion measured per host against the fixed build for its SKU rather than by 'approved' or 'downloaded' in the management console. The reboot is the load-bearing detail: the packet records no live-patch mechanism and states the update requires a reboot to load the fixed clfs.sys, so a host that has installed the update and not restarted is still executing the vulnerable driver and must be counted as exposed. On servers and always-on workstations that pending reboot is precisely the step deferred to protect uptime, and a deferral recorded as 'patched' is the specific way this remediation goes wrong. There is also no lever to trade against the update here, which is what separates this entry from a vulnerable third-party driver: the packet names no compensating configuration, no live-patch path and no way to prevent the driver loading, so a driver-blocklist or load-prevention posture is not an interim mitigation for this flaw — the update that replaces clfs.sys is the whole remediation. Note finally what the control cannot be substituted with: the attacker in this packet is already executing locally as a low-privileged user, so tightening account privilege does not contain the escalation. That is the entire value of the bug to an intruder — it converts access he already holds into SYSTEM — and it is why account-hardening evidence reads clean on a host that is fully exploitable. Priority follows the packet rather than the CVSS band: 7.8 is a local-privilege score, but a public PoC plus confirmed in-the-wild exploitation put the RWEP at 70, and a flaw of this class is the escalation link in a chain rather than a standalone endpoint item.",
+        "evidence": "Packet attack_vector: 'A local attacker crafts a malicious CLFS base log (.BLF) file to trigger a heap overflow / out-of-bounds write in clfs.sys, corrupting kernel memory to elevate a low-privileged process to SYSTEM.' CWE-122 and CWE-787. patch_available true; live_patch_available false with live_patch_notes 'No vendor live-patch mechanism; remediation is the April 2023 Windows cumulative security update, which requires a reboot to load the fixed clfs.sys.' CISA KEV-listed 2023-04-11, active_exploitation confirmed, poc_available true, rwep_score 70 against cvss 7.8. Citing gaps AU-Essential-8-Patch, NIS2-Art21-patch-management, NIST-800-53-SI-2, ISO-27001-2022-A.8.8.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "NIS2-Art21-patch-management",
+          "NIST-800-53-SI-2",
+          "ISO-27001-2022-A.8.8"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-003",
+        "name": "KERNEL-EXPLOITATION-DETECTION",
+        "description": "Between the 2023-04-11 KEV listing and the reboot that actually loads the fixed clfs.sys, detection is the only lever an operator holds, and the packet describes the exploitation behaviour precisely enough to key a rule on it rather than on an assumed signal. The rule needs both halves the packet names. First the input: a CLFS base log — a .BLF file — being created or written by a process that is not running with administrative rights, and outside the locations where the operating system's own components keep their CLFS logs. Second the outcome: a low-privileged process becoming SYSTEM, observable as a SYSTEM-integrity process whose parent was running as a standard user, or as a privilege transition on a process that started unprivileged. Neither half stands alone — legitimate software does write CLFS logs, and SYSTEM processes are created constantly by the service control manager — and it is the pairing inside a short window, the same standard-user context supplying a base-log file and then holding SYSTEM, that distinguishes this exploit from either in isolation. Be clear about what will not see it: nothing is installed, no third-party driver is loaded, no service is created, and the process need not crash, so file-integrity monitoring, driver-load auditing and signature-based endpoint tooling have no artifact to match, and a rule that alerts on kernel crashes or on named exploit tooling would miss an attempt behaving exactly as the packet describes. Preconditions, stated rather than assumed: this requires kernel-level file-creation, process-creation and token telemetry already being collected and shipped off-host before the attempt — a rule authored afterwards against telemetry nobody was recording produces nothing — and detection does not prevent the escalation. Once it succeeds the attacker holds SYSTEM, so the alert buys response time and nothing more; a host that fired this rule goes to incident response and credential rotation, and is not closed out by the cumulative update that arrives later.",
+        "evidence": "The behaviour keyed on is the packet's own attack_vector: 'A local attacker crafts a malicious CLFS base log (.BLF) file to trigger a heap overflow / out-of-bounds write in clfs.sys, corrupting kernel memory to elevate a low-privileged process to SYSTEM' — the .BLF artifact and the low-privileged-to-SYSTEM transition are both stated there. poc_available true and active_exploitation confirmed, so attempts against unremediated hosts are the packet's stated condition. The detection window exists because live_patch_available is false and live_patch_notes require 'the April 2023 Windows cumulative security update, which requires a reboot to load the fixed clfs.sys' — remediation cannot be instantaneous across an estate. Citing gap UK-CAF-B4 (System security).",
+        "gap_closes": [
+          "UK-CAF-B4"
+        ]
+      }
+    ]
   },
   "CVE-2023-28205": {
     "name": "Apple Multiple Products WebKit Use-After-Free Vulnerability (CVE-2023-28205)",
@@ -59677,7 +60637,30 @@
         "adequate": false,
         "gap": "A.8.8 technical-vulnerability management depends on the tool being inventoried and vendor-tracked; Cobalt Strike clients are often ad-hoc installs, so the incomplete-fix RCE was easy to miss until Fortra shipped 4.7.2 and CISA listed it."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-055",
+        "name": "SECURITY-TOOL-INTEGRITY-VERIFICATION",
+        "description": "Cobalt Strike is security tooling that a team installs and runs for itself, which is exactly why every patch-shaped control cited on this entry can pass while a 4.7.1 client keeps rendering implant data: the client is a per-operator install that managed software distribution typically never sees, so the estate's patch report has no row for it and its absence reads as compliance rather than as a gap. The requirement here is that the Cobalt Strike client be carried in the same vulnerability-management inventory and on the same SLA as any other privileged software, counted per operator workstation and not per team server -- the packet's fix is client-side, Fortra shipped 4.7.2 out of band to escape HTML in the Swing components, and remediation is upgrading that client and restarting it, with no live-patch path available. The trust-anchor-inversion test this control demands is available verbatim from the behaviour the packet documents: in a lab, have a controlled implant return data containing an HTML <object> tag, display it in the console, and confirm it renders as inert text instead of instantiating a Java object and calling its setters. Re-run that after every client rebuild, because a re-imaged operator workstation is where a 4.7.1 JAR comes back onto a team that had finished remediating. Precondition: the escaping is a property of the 4.7.2 build. This control is the inventory, the SLA and the test that establish each client actually reached it -- it gives nothing to an operator still running 4.7.1, and a console that displayed attacker-controlled implant data before the upgrade already executed whatever it was sent, so that workstation and the credentials it held belong on the incident path rather than on the patch list.",
+        "evidence": "Packet vector: 'Cobalt Strike 4.7.1 fails to properly escape HTML tags when they are displayed on Swing components. By injecting crafted HTML code, it is possible to remotely execute code in the Cobalt Strike UI.' live_patch_notes: 'No live-patch mechanism; Fortra released Cobalt Strike 4.7.2 as an out-of-band fix that properly escapes HTML in Swing components - remediation is upgrading the client and restarting it.' cisa_kev true, kev_date 2023-03-30, active_exploitation 'confirmed', poc_available true, cvss 9.8, rwep_score 62, patch_available true, live_patch_available false.",
+        "gap_closes": [
+          "AU-Essential-8-Patch",
+          "ISO-27001-2022-A.8.8",
+          "NIST-800-53-SI-2",
+          "NIS2-Art21-patch-management"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-046",
+        "name": "PEN-TEST-SCOPE-INCLUDES-SECURITY-PRODUCTS",
+        "description": "This CVE runs the engagement's data flow backwards, and that is the part scoping documents do not anticipate. The packet has attacker-controlled implant data carrying an HTML <object> tag rendered by the Cobalt Strike client's Java Swing UI, which instantiates arbitrary Java objects and calls their setters, executing code in the operator's console -- so every host the team implants is an input source into the team's own console, and the console is reachable by whoever controls what that host returns. Scope and threat-model language must therefore name the engagement toolchain itself, the team server and each operator client, as attack surface with an untrusted input path, with the client's rendering of implant-returned data enumerated as the specific surface rather than left as instrumentation assumed to sit outside the tested boundary. Keep the scope to those components: the packet ties the defect to the Cobalt Strike client's Swing rendering and gives no basis for treating other consoles or agents in the estate as instances of it. Precondition, and it is the whole limit of this control: scoping and testing describe the surface, they do not remove it. On a team still running the affected client the only closure is the packet's fix -- Cobalt Strike 4.7.2, client upgraded and restarted, with no live-patch path -- and until that lands, any implant on a host the team does not fully control is a live path into the operator's console regardless of how the engagement is scoped.",
+        "evidence": "Packet attack_vector: 'Attacker-controlled implant data containing an HTML <object> tag is rendered by the Cobalt Strike client's Java Swing UI, which instantiates arbitrary Java objects and calls their setters, executing code in the operator's console.' cwe_refs CWE-116. live_patch_notes gives the fix as Cobalt Strike 4.7.2 with the client upgraded and restarted; live_patch_available false. cisa_kev true, kev_date 2023-03-30, active_exploitation 'confirmed', poc_available true. Citing gaps include UK-CAF-B4 (System security).",
+        "gap_closes": [
+          "UK-CAF-B4"
+        ]
+      }
+    ]
   },
   "CVE-2022-39197": {
     "name": "Fortra Cobalt Strike Teamserver Cross-Site Scripting (XSS) Vulnerability",
@@ -60226,7 +61209,30 @@
         "adequate": false,
         "gap": "A.8.7 anti-malware posture that depends on OS reputation warnings is defeated by the fail-open MOTW bypass; the control passes the ransomware installer without an alert."
       }
-    }
+    },
+    "new_control_requirements": [
+      {
+        "id": "NEW-CTRL-041",
+        "name": "PROTECTION-MECHANISM-FAILURE-CLASS-REGRESSION",
+        "description": "SmartScreen is exactly the protection-mechanism class this control governs, and the packet describes the mechanism failing open rather than an application being exploited: an MSI or PE carrying a malformed Authenticode signature is delivered by drive-by download, SmartScreen errors while parsing that signature, the Mark-of-the-Web reputation check is skipped, and the file runs with no warning and installs Magniber ransomware. Bound to this estate, the control means every layer credited with covering the MOTW/SmartScreen class — the detonation chamber, the EDR rules written against 'unknown executable downloaded from the internet', and any AppLocker or WDAC policy an attestation leans on — carries a regression battery containing this primitive and runs it on every patch deployment, not once at remediation. The distinguishing test keys on the behaviour the packet documents rather than on a malware sample or a policy read-back: put an MSI and a PE with a deliberately malformed Authenticode signature and a Mark-of-the-Web tag on a test host, open them the way a drive-by download would, and confirm the reputation check actually runs and the file is warned on or blocked. Confirming that SmartScreen is 'enabled' in policy proves nothing here — the setting was enabled on every host this CVE was exploited against; the bypass is the check being skipped, not turned off. Precondition: this is a test regime and not a mitigation. It tells the operator whether the class is genuinely covered on a given build; it does not close the bypass. The packet records the March 2023 Windows security update as what makes SmartScreen fail closed on malformed signatures, with no live-patch path and a reboot required, so on a host that has not restarted the battery will still show the file executing silently — which is the correct result, and means that host is exposed rather than that the test failed.",
+        "evidence": "Packet fields for CVE-2023-24880 (Microsoft Windows SmartScreen Security Feature Bypass Vulnerability, CWE-863): CISA KEV listed 2023-03-14, active_exploitation confirmed, poc_available true, CVSS 4.4, rwep_score 66. The attack_vector states the attacker crafts an MSI/PE with a malformed Authenticode signature and delivers it via drive-by download; when the target opens it, SmartScreen errors while parsing the signature and skips the Mark-of-the-Web reputation check, so the file runs without the usual warning and installs Magniber ransomware. patch_available true, live_patch_available false, with live_patch_notes stating remediation requires the March 2023 Windows security update that makes SmartScreen fail closed on malformed signatures, followed by a reboot. The entry's citing gaps are the malicious-code-protection, anti-malware and user-application-hardening controls that credit SmartScreen as the covering mechanism.",
+        "gap_closes": [
+          "NIST-800-53-SI-3",
+          "ISO-27001-2022-A.8.7",
+          "AU-Essential-8-App-Hardening"
+        ]
+      },
+      {
+        "id": "NEW-CTRL-001",
+        "name": "CISA-KEV-RESPONSE-SLA",
+        "description": "This entry is the case the control exists for, because its severity band and its real-world priority point in opposite directions: the packet records CVSS 4.4 — below the threshold at which most patch policies open an expedited clock — against confirmed in-the-wild exploitation, a public PoC, a KEV listing dated 2023-03-14 and an RWEP of 66, with the observed outcome being ransomware installation. A programme that schedules by CVSS band puts this update in the routine monthly stream while the exploitation it is listed for is already running. For this CVE the requirement is that the clock starts at the KEV listing and that completion is measured per host as the March 2023 Windows security update installed and the machine restarted, not as 'approved', 'downloaded' or 'pending reboot' in the management console — the packet registers no live-patch path and states the fix takes effect only after that reboot, so a host holding the update without a restart is still running the code that skips the reputation check. Precondition, stated rather than glossed: the packet names no interim vendor mitigation for this entry, so there is nothing to deploy in place of the update. During the window before each host restarts, the only cover is on the delivery and execution side — which is a different control, and it must not be recorded as though it satisfied this one. Nor does the SLA reach a host already compromised: the packet's outcome is ransomware installed by a file that ran silently, and installing the update afterwards does not undo an execution that already happened.",
+        "evidence": "Packet fields for CVE-2023-24880: cisa_kev true with kev_date 2023-03-14, active_exploitation confirmed, poc_available true, cvss 4.4, rwep_score 66. patch_available true, live_patch_available false, live_patch_notes stating remediation requires the March 2023 Windows security update that makes SmartScreen fail closed on malformed signatures, followed by a reboot. The attack_vector records the exploitation outcome as the crafted MSI/PE running without the usual warning and installing Magniber ransomware. The entry's citing gaps include the EU NIS2 vulnerability-handling/patch-management control and the UK CAF system-security control, both of which an estate can attest while scheduling this CVE by its 4.4 severity band.",
+        "gap_closes": [
+          "NIS2-Art21-patch-management",
+          "UK-CAF-B4"
+        ]
+      }
+    ]
   },
   "CVE-2022-41328": {
     "name": "Fortinet FortiOS Path Traversal Vulnerability",